Computed tomography multi-planar and 3D image assessment protocol for detailed analysis of inner ear malformations in patients undergoing cochlear implantation counseling.

<b><br>Introduction:</b> Congenital inner ear malformations resulting from embryogenesis may be visualized in radiological scans. Many attempts have been made to describe and classify the defects of the inner ear based on anatomical and radiological findings.</br> <b><br>Aim:</b> The aim was to propose and discuss computed tomography multi-planar and 3D image assessment protocols for detailed analysis of inner ear malformations in patients undergoing cochlear implantation counseling.</br> <b><br>Material and methods:</b> A retrospective analysis of 22 malformed inner ears. CT scans were analyzed using the Multi-Planar Reconstruction (MPR) option and 3D reconstruction.</br> <b><br>Results:</b> The protocol of image interpretation was developed to allow reproducibility for evaluating each set of images. The following malformations were identified: common cavity, cochlear hypoplasia type II, III, and IV, incomplete partition type II and III, and various combinations of vestibule labyrinth malformations. All anomalies have been presented and highlighted in figures with appropriate descriptions for easier identification. Figures of normal inner ears were also included for comparison. 3D reconstructions for each malformation were presented, adding clinical value to the detailed analysis.</br> <b><br>Conclusions:</b> Properly analyzing CT scans in cochlear implantation counseling is a necessary and beneficial tool for appropriate candidate selection and preparation for surgery. As proposed in this study, the unified scans evaluation scheme simplifies the identification of malformations and reduces the risk of omitting particular anomalies. Multi-planar assessment of scans provides most of the necessary details. The 3D reconstruction technique is valuable in addition to diagnostics influencing the decision-making process. It can minimize the risk of misdiagnosis. Disclosure of the inner ear defect and its precise imaging provides detailed anatomical knowledge of each ear, enabling the selection of the appropriate cochlear implant electrode and the optimal surgical technique.</br>.

RT-SRTS: Angle-agnostic real-time simultaneous 3D reconstruction and tumor segmentation from single X-ray projection.

Radiotherapy is one of the primary treatment methods for tumors, but the organ movement caused by respiration limits its accuracy. Recently, 3D imaging from a single X-ray projection has received extensive attention as a promising approach to address this issue. However, current methods can only reconstruct 3D images without directly locating the tumor and are only validated for fixed-angle imaging, which fails to fully meet the requirements of motion control in radiotherapy. In this study, a novel imaging method RT-SRTS is proposed which integrates 3D imaging and tumor segmentation into one network based on multi-task learning (MTL) and achieves real-time simultaneous 3D reconstruction and tumor segmentation from a single X-ray projection at any angle. Furthermore, the attention enhanced calibrator (AEC) and uncertain-region elaboration (URE) modules have been proposed to aid feature extraction and improve segmentation accuracy. The proposed method was evaluated on fifteen patient cases and compared with three state-of-the-art methods. It not only delivers superior 3D reconstruction but also demonstrates commendable tumor segmentation results. Simultaneous reconstruction and segmentation can be completed in approximately 70 ms, significantly faster than the required time threshold for real-time tumor tracking. The efficacies of both AEC and URE have also been validated in ablation studies. The code of work is available at https://github.com/ZywooSimple/RT-SRTS.

Craniovertebral Junction Surgical Approaches: State of Art.

Surgical approaches directed toward craniovertebral junction (CVJ) can be addressed to the ventral, dorsal, and lateral aspects through a variety of 360° surgical corridors Herein, we report features, advantages, and limits of the updated technical support in CVJ surgery in clinical setting and dissection laboratories enriched by our preliminary surgical results of the simultaneous application of O-arm intraoperative neuronavigation and imaging system along with the 3D-4K EX in TOA for the treatment of CVJ pathologies.In the past 4 years, eight patients harboring CVJ compressive pathologies underwent one-step combined anterior neurosurgical decompression and posterior instrumentation and fusion technique with the aid of exoscope and O-arm. In our equipped Cranio-Vertebral Junction Laboratory, we use fresh cadavers (and injected "head and neck" specimens) whose policy, protocols, and logistics have already been elucidated in previous works. Five fresh-frozen adult specimens were dissected adopting an FLA. In these specimens, a TOA was also performed, as well as a neuronavigation-assisted comparison between transoral and transnasal explorable distances.A complete decompression along with stable instrumentation and fusion of the CVJ was accomplished in all the cases at the maximum follow-up (mean: 25.3 months). In two cases, the O-arm navigation allowed the identification of residual compression that was not clearly visible using the microscope alone. In four cases, it was not possible to navigate C1 lateral masses and C2 isthmi due to the angled projection unfitting with the neuronavigation optical system, so misleading the surgeon and strongly suggesting changing surgical strategy intraoperatively. In another case (case 4), it was possible to navigate and perform both C1 lateral masses and C2 isthmi screwing, but the screw placement was suboptimal at the immediate postoperative radiological assessment. In this case, the hardware displacement occurred 2 months later requiring reoperation.

Reliability and validity of handheld structured light scanners and a static stereophotogrammetry system in facial three-dimensional surface imaging.

Several new systems for three-dimensional (3D) surface imaging of the face have become available to assess changes following orthognathic or facial surgery. Before they can be implemented in practice, their reliability and validity must be established. Our aim, therefore, was to study the intra- and inter-system reliability and validity of 3dMD (stereophotogrammetry), Artec Eva and Artec Space Spider (both structured light scanners). Intra- and inter-system reliability, expressed in root mean square distance, was determined by scanning a mannequin's head and the faces of healthy volunteers multiple times. Validity was determined by comparing the linear measurements of the scans with the known distances of a 3D printed model. Post-processing errors were also calculated. Intra-system reliability after scanning the mannequin's head was best with the Artec Space Spider (0.04 mm Spider; 0.07 mm 3dMD; 0.08 mm Eva). The least difference in inter-system reliability after scanning the mannequin's head was between the Artec Space Spider and Artec Eva. The best intra-system reliability after scanning human subjects was with the Artec Space Spider (0.15 mm Spider; 0.20 mm Eva; 0.23 mm 3dMD). The least difference in inter-system reliability after scanning human subjects was between the Artec Eva and Artec Space Spider. The most accurate linear measurement validity occurred with the Artec Space Spider. The post-processing error was 0.01 mm for all the systems. The Artec Space Spider is the most reliable and valid scanning system.

Noninvasive virtual biopsy using micro-registered optical coherence tomography (OCT) in human subjects.

Histological hematoxylin and eosin-stained (H&E) tissue sections are used as the gold standard for pathologic detection of cancer, tumor margin detection, and disease diagnosis. Producing H&E sections, however, is invasive and time-consuming. While deep learning has shown promise in virtual staining of unstained tissue slides, true virtual biopsy requires staining of images taken from intact tissue. In this work, we developed a micron-accuracy coregistration method [micro-registered optical coherence tomography (OCT)] that can take a two-dimensional (2D) H&E slide and find the exact corresponding section in a 3D OCT image taken from the original fresh tissue. We trained a conditional generative adversarial network using the paired dataset and showed high-fidelity conversion of noninvasive OCT images to virtually stained H&E slices in both 2D and 3D. Applying these trained neural networks to in vivo OCT images should enable physicians to readily incorporate OCT imaging into their clinical practice, reducing the number of unnecessary biopsy procedures.

Preoperative three-dimensional modelling and virtual reality planning aids nephron sparing surgery in a child with bilateral Wilms tumour.

Bilateral Wilms tumour (BWT) is a surgically challenging condition. Virtual reality (VR) reconstruction aids surgeons to foresee the anatomy ahead of Nephron Sparing Surgery (NSS). Three-dimensional (3D) visualisation improves the anatomical orientation of surgeons performing NSS. We herewith report a case of BWT where VR planning and 3D printing were used to aid NSS. Conventional imaging is often found to be inadequate while assessing the tumour-organ-vascular anatomy. Advances like VR and 3D printing help surgeons plan better for complex surgeries like bilateral NSS. Next-generation extended reality tools will likely aid robotic-assisted precision NSS and improve patient outcomes.

Laparoscopy training of novices with complex curved instruments using 2D- and 3D-visualization.

PURPOSE: Beside many advantages, disadvantages such as reduced degrees of freedom and poorer depth perception are still apparent in laparoscopic surgery. 3D visualization and the development of complex instruments are intended to counteract the disadvantages. We want to find out whether the use of complex instruments and 3D visualization has an influence on the performance of novices. METHODS: 48 medical students with no experience in laparoscopic surgery or simulator-based laparoscopy training were included. They were randomized in four groups according to a stratification assessment. During a structured training period they completed the FLS-Tasks "PEG Transfer", "Pattern Cut" and "Intracorporeal Suture" and a transfer task based on these three. Two groups used conventional laparoscopic instruments with 3D or 2D visualization, two groups used complex curved instruments. The groups were compared in terms of their performance. RESULTS: In 2D laparoscopy there was a better performance with straight instruments vs. curved instruments in PEG Transfer and Intracorporeal Suture. In the transfer task, fewer errors were made with straight instruments. In 2D vs. 3D laparoscopy when using complex curved instruments there was an advantage in Intracorporeal Suture and PEG Transfer for 3D visualization. Regarding the transfer exercise, a better performance was observed and fewer errors were made in 3D group. CONCLUSION: We could show that learning laparoscopic techniques with complex curved instruments is more difficult with standard 2D visualization and can be overcome using 3D optics. The use of curved instruments under 3D vision seems to be advantageous when working on more difficult tasks.

Robot-assisted anterior cruciate ligament reconstruction based on three-dimensional images.

Background Tunnel placement is a key step in anterior cruciate ligament (ACL) reconstruction. The purpose of this study was to evaluate the accuracy of bone tunnel drilling in arthroscopic ACL reconstruction assisted by a three-dimensional (3D) image-based robot system. Methods Robot-assisted ACL reconstruction was performed on twelve freshly frozen knee specimens. During the operation, three-dimensional images were used for ACL bone tunnel planning, and the robotic arm was used for navigation and drilling. Twelve patients who underwent traditional arthroscopic ACL reconstruction were included. 3D computed tomography was used to measure the actual position of the ACL bone tunnel and to evaluate the accuracy of the robotic and traditional ACL bone tunnel. Results On the femoral side, the positions of robotic and traditional surgery tunnels were 29.3 ± 1.4% and 32.1 ± 3.9% in the deep-to-shallow direction of the lateral femoral condyle (p = 0.032), and 34.6 ± 1.2% and 21.2 ± 9.4% in the high-to-low direction (p < 0.001), respectively. On the tibial side, the positions of the robotic and traditional surgical tunnels were located at 48.4 ± 0.9% and 45.8 ± 2.8% of the medial-to-lateral diameter of the tibial plateau (p = 0.008), 38.1 ± 0.8% and 34.6 ± 6.0% of the anterior-to-posterior diameter (p = 0.071), respectively. Conclusions In this study, ACL reconstruction was completed with the assistance of a robot arm and 3D images, and the robot was able to drill the bone tunnel more accurately than the traditional arthroscopic ACL reconstruction.

An evaluation of the effectiveness of 'ULTRACOL 200' in enhancing nasolabial fold wrinkles through cutaneous repair.

BACKGROUND: Injectable filler, a nonsurgical beauty method, has gained popularity in rejuvenating sagging skin. In this study, polydioxanone (PDO) was utilized as the main component of the ULTRACOL200 filler that helps stimulate collagenesis and provide skin radiant effects. The study aimed to evaluate and compare the effectiveness of ULTRACOL200 with other commercialized products in visually improving dermatological problems. METHODS: Herein, 31 participants aged between 20 and 59 years were enrolled in the study. 1 mL of the testing product, as well as the quantity for the compared groups was injected into each participants face side individually. Subsequently, skin texture and sunken volume of skin were measured using ANTERA 3D CS imaging technology at three periods: before the application, 4 weeks after the initial application, and 4 weeks after the 2nd application of ULTRACOL200. RESULTS: The final results of skin texture and wrinkle volume evaluation consistently demonstrated significant enhancement. Consequently, subjective questionnaires were provided to the participants to evaluate the efficacy of the testing product, illustrating satisfactory responses after the twice applications. CONCLUSION: The investigation has contributed substantially to the comprehension of a PDO-based filler (ULTRACOL200) for skin enhancement and provided profound insight for future clinical trials.

[Spatial distribution pattern of local tumor progression analysis after microwave ablation of hepatocellular carcinoma based on three-dimensional magnetic resonance imaging].

Objective: To investigate the spatial distribution pattern of local tumor progression (LTP) for hepatocellular carcinoma (HCC) ≤5 cm after microwave ablation. Methods: A retrospective analysis was performed on 169 HCCs with matched MRI before and after ablation from December 2009 to December 2019. A tumor MRI was reconstructed using three-dimensional visualization technology. LTP was classified as contact or non-contact, early or late stage, according to whether LTP was in contact with the edge of the ablation zone and the occurrence time (24 months). The tumor-surrounded area was divided into eight quadrants by using the eight-quadrant map method. An analysis was conducted on the spatial correlation between the quadrant where the ablative margin (AM) safety boundary was located and the quadrant where different types of LTP occurred. The t-test, or rank-sum test, was used for the measurement data. 2-test for count data was used to compare the difference between the two groups. Results: The AM quadrant had a distribution of 54.4% LTP, 64.2% early LTP stage, and 69.1% contact LTP, suggesting this quadrant was much more concentrated than the other quadrants (P < 0.001). Additionally, the AM quadrant had only 15.2% of non-contact type LTP and 17.1% of late LTP, which was not significantly different from the average distribution probability of 12.5% (100/8%) among the eight quadrants (P = 0.667, 0.743). 46.6% of early contact type LTP was located at the ablation needle tip, 25.2% at the body, and 28.1% at the caudal, while the location distribution probabilities of non-early contact LTP were 34.8%, 31.8%, and 33.3%, respectively. Conclusion: LTP mostly occurs in areas where the ablation safety boundary is the shortest. However, non-contact LTP and late LTP stages exhibit the feature of uniform distribution. Thus, this type of LPT may result from an inadequate non-ablation safety boundary.

Delineation of three-dimensional tumor margins based on normalized absolute difference mapping via volumetric optical coherence tomography.

The extent of surgical resection is an important prognostic factor in the treatment of patients with glioblastoma. Optical coherence tomography (OCT) imaging is one of the adjunctive methods available to achieve the maximal surgical resection. In this study, the tumor margins were visualized with the OCT image obtained from a murine glioma model. A commercialized human glioblastoma cell line (U-87) was employed to develop the orthotopic murine glioma model. A swept-source OCT (SS-OCT) system of 1300 nm was used for three-dimensional imaging. Based on the OCT intensity signal, which was obtained via accumulation of each A-scan data, an en-face optical attenuation coefficient (OAC) map was drawn. Due to the limited working distance of the focused beam, OAC values decrease with depth, and using the OAC difference in the superficial area was chosen to outline the tumor boundary, presenting a challenge in analyzing the tumor margin along the depth direction. To overcome this and enable three-dimensional tumor margin detection, we converted the en-face OAC map into an en-face difference map with x- and y-directions and computed the normalized absolute difference (NAD) at each depth to construct a volumetric NAD map, which was compared with the corresponding H&E-stained image. The proposed method successfully revealed the tumor margin along the peripheral boundaries as well as the margin depth. We believe this method can serve as a useful adjunct in glioma surgery, with further studies necessary for real-world practical applications.

Influence of augmented reality technique on the accuracy of autotransplanted teeth in surgically created sockets.

BACKGROUND: The objective of the present study was to evaluate the reliability of an augmented reality drilling approach and a freehand drilling technique for the autotransplantation of single-rooted teeth. MATERIALS AND METHODS: Forty samples were assigned to the following surgical techniques for drilling guidance of the artificial sockets: A. augmented reality technique (AR) (n = 20) and B. conventional free-hand technique (FT) (n = 20). Then, two models with 10 teeth each were submitted to a preoperative cone-beam computed tomography (CBCT) scan and a digital impression by a 3D intraoral scan. Afterwards, the autotrasplanted teeth were planned in a 3D dental implant planning software and transferred to the augmented reality device. Then, a postoperative CBCT scan was performed. Data sets from postoperative CBCT scans were aligned to the planning in the 3D implant planning software to analize the coronal, apical and angular deviations. Student's t-test and Mann-Whitney non-parametric statistical analysis were used to analyze the results. RESULTS: No statistically significant differences were shown at coronal (p = 0.123) and angular (p = 0.340) level; however, apical deviations between AR and FT study groups (p = 0.008) were statistically significant different. CONCLUSION: The augmented reality appliance provides higher accuracy in the positioning of single-root autotransplanted teeth compared to the conventional free-hand technique.

Precisão dos escaneamentos intraorais e modelos 3D, com diferentes escâneres e softwares: revisão sistemática / Accuracy of intraoral scanning and 3D models, with different scanners and software: systematic review

Objetivo: Esta revisão sistemática investigou se a precisão dos escaneamentos intraorais e modelos 3D com diferentes escâneres e softwares, é suficiente para os clínicos aceitarem e implementarem as novas tecnologias como critério de diagnóstico e planejamentos dos tratamentos. Material e métodos: Protocolo PROSPERO número CRD42020218151. Duas revisoras realizaram uma pesquisa avançada de banco de dados eletrônico, sem restrição de idioma ou data, no MEDLINE/PubMed; Embase; BVS/LILACS; Scopus; Cochrane Library; Google Scholar e Web of Science até janeiro de 2021. Os estudos foram escolhidos por título e resumo para triagem, de acordo com os seguintes critérios de inclusão: Estudos Clínicos in vivo e in vitro (pacientes, modelos de gesso de pacientes e manequins simulando bocas humanas) com os diferentes softwares e técnicas de escaneamento intraoral comparando a acurácia, fidelidade e/ou precisão como desfecho dos escaneamentos intraorais, dos modelos virtuais em 3D com modelos de gesso; com um mínimo de 5 pacientes escaneados sem limite de idade ou gênero; texto completo acessível; nos estudos de coorte, transversais e caso-controle. Após a leitura do texto completo os artigos foram excluídos de acordo com os seguintes critérios: 1) estudos com menos de 5 pacientes; 2) estudos em animais; 3) revisão sistemática, descrição de técnica, intervenções, protocolos; 4) escaneamentos por Ressonância Magnética ou Tomografia computadorizada; 5) estudos em que não foram utilizados scanners intraorais. Resultados. Dos 4410 estudos inicialmente identificados, 16 preencheram os critérios de inclusão. O guia PRISMA foi utilizado para redação da revisão e a ferramenta ROBINS-1 da Cochrane foi utilizada para análise de viés. Os estudos incluídos na sua maioria mostraram diferenças estatisticamente significativas entre os resultados das moldagens convencionais e digitais, porém também demonstraram que as diferenças não têm significância clínica. Conclusões: Esta revisão sistemática permitiu levantar dados que demonstraram que os escaneamentos intraorais não são superiores às moldagens convencionais, mas de acurácia equivalente com confiabilidade para o uso das imagens digitais conseguidas por escaneamento intraoral e dos modelos digitais provenientes destes escaneamentos.

Accuracy of implant placement with computer-aided static, dynamic, and robot-assisted surgery: a systematic review and meta-analysis of clinical trials.

This systematic review explores the accuracy of computerized guided implant placement including computer-aided static, dynamic, and robot-assisted surgery. An electronic search up to February 28, 2023, was conducted using the PubMed, Embase, and Scopus databases using the search terms "surgery", "computer-assisted", "dynamic computer-assisted", "robotic surgical procedures", and "dental implants". The outcome variables were discrepancies including the implant's 3D-coronal, -apical and -angular deviations. Articles were selectively retrieved according to the inclusion and exclusion criteria, and the data were quantitatively meta-analysed to verify the study outcomes. Sixty-seven articles were finally identified and included for analysis. The accuracy comparison revealed an overall mean deviation at the entry point of 1.11 mm (95% CI: 1.02-1.19), and 1.40 mm (95% CI: 1.31-1.49) at the apex, and the angulation was 3.51˚ (95% CI: 3.27-3.75). Amongst computerized guided implant placements, the robotic system tended to show the lowest deviation (0.81 mm in coronal deviation, 0.77 mm in apical deviation, and 1.71˚ in angular deviation). No significant differences were found between the arch type and flap operation in cases of dynamic navigation. The fully-guided protocol demonstrated a significantly higher level of accuracy compared to the pilot-guided protocol, but did not show any significant difference when compared to the partially guided protocol. The use of computerized technology clinically affirms that operators can accurately place implants in three directions. Several studies agree that a fully guided protocol is the gold standard in clinical practice.

Development of a modified 3D region proposal network for lung nodule detection in computed tomography scans: a secondary analysis of lung nodule datasets.

BACKGROUND: Low-dose computed tomography (LDCT) has been shown useful in early lung cancer detection. This study aimed to develop a novel deep learning model for detecting pulmonary nodules on chest LDCT images. METHODS: In this secondary analysis, three lung nodule datasets, including Lung Nodule Analysis 2016 (LUNA16), Lung Nodule Received Operation (LNOP), and Lung Nodule in Health Examination (LNHE), were used to train and test deep learning models. The 3D region proposal network (RPN) was modified via a series of pruning experiments for better predictive performance. The performance of each modified deep leaning model was evaluated based on sensitivity and competition performance metric (CPM). Furthermore, the performance of the modified 3D RPN trained on three datasets was evaluated by 10-fold cross validation. Temporal validation was conducted to assess the reliability of the modified 3D RPN for detecting lung nodules. RESULTS: The results of pruning experiments indicated that the modified 3D RPN composed of the Cross Stage Partial Network (CSPNet) approach to Residual Network (ResNet) Xt (CSP-ResNeXt) module, feature pyramid network (FPN), nearest anchor method, and post-processing masking, had the optimal predictive performance with a CPM of 92.2%. The modified 3D RPN trained on the LUNA16 dataset had the highest CPM (90.1%), followed by the LNOP dataset (CPM: 74.1%) and the LNHE dataset (CPM: 70.2%). When the modified 3D RPN trained and tested on the same datasets, the sensitivities were 94.6%, 84.8%, and 79.7% for LUNA16, LNOP, and LNHE, respectively. The temporal validation analysis revealed that the modified 3D RPN tested on LNOP test set achieved a CPM of 71.6% and a sensitivity of 85.7%, and the modified 3D RPN tested on LNHE test set had a CPM of 71.7% and a sensitivity of 83.5%. CONCLUSION: A modified 3D RPN for detecting lung nodules on LDCT scans was designed and validated, which may serve as a computer-aided diagnosis system to facilitate lung nodule detection and lung cancer diagnosis.

A modified 3D RPN for detecting lung nodules on CT images that exhibited greater sensitivity and CPM than did several previously reported CAD detection models was established.

[Educational Impact of Using Three-dimentional Models for Preoperative Simulation of Segmentectomy for Medical Students].

BACKGROUND: Lung segment anatomy is complex and difficult for medical students to comprehend, so education is a challenging issue. Three-dimensional (3D) models may be a more effective teaching tool for medical students. We evaluated the usefulness of 3D models in education for segmentectomy. METHODS: Sixty-six students who participated in clinical training for thoracic surgery from November 2020 to March 2022 were included in the study. Computed tomography (CT) images, 3D-CT images, and 3D models were used to investigate students' level of understanding for each lung segment. Levels of understanding were assessed using a questionnaire administered using the interview method. The results of the questionnaire were tabulated and analyzed using the Kruskal-Wallis test. RESULTS: The response rate was 100%. Comprehension scores for all questions were highest for 3D models, followed by 3D-CT and CT. These results suggest that use of a 3D entity model may have a high educational effect. CONCLUSION: Although it is unclear whether students' comprehension was directly related to their understanding of correct anatomy, our results suggest that 3D models may be an effective way for medical students to understand lung segmentectomy.

3D cine-magnetic resonance imaging using spatial and temporal implicit neural representation learning (STINR-MR).

Objective. 3D cine-magnetic resonance imaging (cine-MRI) can capture images of the human body volume with high spatial and temporal resolutions to study anatomical dynamics. However, the reconstruction of 3D cine-MRI is challenged by highly under-sampled k-space data in each dynamic (cine) frame, due to the slow speed of MR signal acquisition. We proposed a machine learning-based framework, spatial and temporal implicit neural representation learning (STINR-MR), for accurate 3D cine-MRI reconstruction from highly under-sampled data.Approach. STINR-MR used a joint reconstruction and deformable registration approach to achieve a high acceleration factor for cine volumetric imaging. It addressed the ill-posed spatiotemporal reconstruction problem by solving a reference-frame 3D MR image and a corresponding motion model that deforms the reference frame to each cine frame. The reference-frame 3D MR image was reconstructed as a spatial implicit neural representation (INR) network, which learns the mapping from input 3D spatial coordinates to corresponding MR values. The dynamic motion model was constructed via a temporal INR, as well as basis deformation vector fields (DVFs) extracted from prior/onboard 4D-MRIs using principal component analysis. The learned temporal INR encodes input time points and outputs corresponding weighting factors to combine the basis DVFs into time-resolved motion fields that represent cine-frame-specific dynamics. STINR-MR was evaluated using MR data simulated from the 4D extended cardiac-torso (XCAT) digital phantom, as well as two MR datasets acquired clinically from human subjects. Its reconstruction accuracy was also compared with that of the model-based non-rigid motion estimation method (MR-MOTUS) and a deep learning-based method (TEMPEST).Main results. STINR-MR can reconstruct 3D cine-MR images with high temporal (<100 ms) and spatial (3 mm) resolutions. Compared with MR-MOTUS and TEMPEST, STINR-MR consistently reconstructed images with better image quality and fewer artifacts and achieved superior tumor localization accuracy via the solved dynamic DVFs. For the XCAT study, STINR reconstructed the tumors to a mean ± SD center-of-mass error of 0.9 ± 0.4 mm, compared to 3.4 ± 1.0 mm of the MR-MOTUS method. The high-frame-rate reconstruction capability of STINR-MR allows different irregular motion patterns to be accurately captured.Significance. STINR-MR provides a lightweight and efficient framework for accurate 3D cine-MRI reconstruction. It is a 'one-shot' method that does not require external data for pre-training, allowing it to avoid generalizability issues typically encountered in deep learning-based methods.

Deformable registration of preoperative MR and intraoperative long-length tomosynthesis images for guidance of spine surgery via image synthesis.

PURPOSE: Improved integration and use of preoperative imaging during surgery hold significant potential for enhancing treatment planning and instrument guidance through surgical navigation. Despite its prevalent use in diagnostic settings, MR imaging is rarely used for navigation in spine surgery. This study aims to leverage MR imaging for intraoperative visualization of spine anatomy, particularly in cases where CT imaging is unavailable or when minimizing radiation exposure is essential, such as in pediatric surgery. METHODS: This work presents a method for deformable 3D-2D registration of preoperative MR images with a novel intraoperative long-length tomosynthesis imaging modality (viz., Long-Film [LF]). A conditional generative adversarial network is used to translate MR images to an intermediate bone image suitable for registration, followed by a model-based 3D-2D registration algorithm to deformably map the synthesized images to LF images. The algorithm's performance was evaluated on cadaveric specimens with implanted markers and controlled deformation, and in clinical images of patients undergoing spine surgery as part of a large-scale clinical study on LF imaging. RESULTS: The proposed method yielded a median 2D projection distance error of 2.0 mm (interquartile range [IQR]: 1.1-3.3 mm) and a 3D target registration error of 1.5 mm (IQR: 0.8-2.1 mm) in cadaver studies. Notably, the multi-scale approach exhibited significantly higher accuracy compared to rigid solutions and effectively managed the challenges posed by piecewise rigid spine deformation. The robustness and consistency of the method were evaluated on clinical images, yielding no outliers on vertebrae without surgical instrumentation and 3% outliers on vertebrae with instrumentation. CONCLUSIONS: This work constitutes the first reported approach for deformable MR to LF registration based on deep image synthesis. The proposed framework provides access to the preoperative annotations and planning information during surgery and enables surgical navigation within the context of MR images and/or dual-plane LF images.

Hybrid-supervised deep learning for domain transfer 3D protoacoustic image reconstruction.

Objective. Protoacoustic imaging showed great promise in providing real-time 3D dose verification of proton therapy. However, the limited acquisition angle in protoacoustic imaging induces severe artifacts, which impairs its accuracy for dose verification. In this study, we developed a hybrid-supervised deep learning method for protoacoustic imaging to address the limited view issue.Approach. We proposed a Recon-Enhance two-stage deep learning method. In the Recon-stage, a transformer-based network was developed to reconstruct initial pressure maps from raw acoustic signals. The network is trained in a hybrid-supervised approach, where it is first trained using supervision by the iteratively reconstructed pressure map and then fine-tuned using transfer learning and self-supervision based on the data fidelity constraint. In the enhance-stage, a 3D U-net is applied to further enhance the image quality with supervision from the ground truth pressure map. The final protoacoustic images are then converted to dose for proton verification.Main results. The results evaluated on a dataset of 126 prostate cancer patients achieved an average root mean squared errors (RMSE) of 0.0292, and an average structural similarity index measure (SSIM) of 0.9618, out-performing related start-of-the-art methods. Qualitative results also demonstrated that our approach addressed the limit-view issue with more details reconstructed. Dose verification achieved an average RMSE of 0.018, and an average SSIM of 0.9891. Gamma index evaluation demonstrated a high agreement (94.7% and 95.7% for 1%/3 mm and 1%/5 mm) between the predicted and the ground truth dose maps. Notably, the processing time was reduced to 6 s, demonstrating its feasibility for online 3D dose verification for prostate proton therapy.Significance. Our study achieved start-of-the-art performance in the challenging task of direct reconstruction from radiofrequency signals, demonstrating the great promise of PA imaging as a highly efficient and accurate tool forinvivo3D proton dose verification to minimize the range uncertainties of proton therapy to improve its precision and outcomes.