Automated Segmentation of Lung Regions in 3D CT Scans Using Hybrid Unsupervised-Supervised Models

Nearly 20% of Total knee Arthroplasty patients remain dissatisfied. This is a major concern in twenty-first century arthroplasty practice. Accurate implant sizing is shown to improve the implant survival, knee balance and patient reported outcome. Aim of the current study is to assess the efficacy of pre-operative three-dimensional (3D) CT scan templating in a robot-assisted TKA in predicting the correct implant sizes and alignment. Prospectively collected data in a single center from 30 RA-TKAswas assessed. Inclusion criterion was patients with end stage arthritis (both osteoarthritis and rheumatoid arthritis) undergoing primary TKA. Patients undergoing revision TKA and patients not willing to participate in the study were excluded. Preliminary study of ten patients had indicated almost 100% accuracy in determining the implant size and position. Sample size was estimated to be 28 for 90% reduction in implant size and position error with α error of 0.05 and beta error of 0.20 with power of study being 80. Post-operative radiographs were assessed by an independent observer with respect to implant size and position. The accuracy of femoral and tibial component sizing in the study was compared with the historic control with Chi-squared test. The p value < 0.05 was considered significant. The pre-operative CT scan 3D templating accuracy was 100% (30 out of 30 knees) for femoral component and 96.67% (29 out of 30 knees) for tibial component. The implant position and limb alignment was accurate in 100% of patients. The accuracy of femoral component and tibial component sizing is statistically significant (Chi-squared test, p value 0.0105 and 0.0461, respectively). The study results show the effectiveness of pre-operative 3 D CT scan planning in predicting the implant sizes and implant positioning. This may have a potential to improve the implant longevity, clinical outcomes and patient satisfaction.

IntroductionAccurate glenoid component placement is crucial for anatomic (TSA) or reverse (RSA) total shoulder arthroplasty. Preoperative glenoid assessment by using CT scans with or without planning software seems to be the established method to plan implant positions. MRI scans can also display the glenoid bone for preoperative assessment while reducing radiation exposure. Therefore, the objective of this study was to manually assess the glenoid version and inclination in 2D MRI and CT scans in cases with degenerative shoulder pathologies. The results were compared to those of an automated 3D planning software to validate the imaging modality for preoperative glenoid assessment. MethodsMRI and CT scans of 146 patients (n=41 aTSA; n=105 RSA) were included in this retrospective, single-center study. Glenoid version and inclination were measured manually according to Friedman et al and Maurer et al on CT and MRI scans by two observers. Subsequently, the results were compared to the automated measurements performed by a planning software. A repeated-measures analysis of variance (ANOVA) was performed to compare the measured angles and interobserver and intraobserver reliability was calculated using the intraclass correlation coefficients. The level of significance was set p<0.05. ResultsThe average glenoid inclination measured in CT scans was 7.94°±7.33°, in MRI scans 8.56°±7.34° and in automated planning software 7.87°±7.60°. The ANOVA analysis revealed significant differences in mean inclination between 2D MRI and 2D CT (p<0.0005) and between MRI and automated software (p=0.011). No significant difference was found between 2D CT scans and automated planning software (p=1.000). Mean glenoid version measured in 2D CT scans was -7.94°±10.86°, in 2D MRI scans it was -8.04°±10.80° and -8.32°±11.53° by the automated planning software. There was no significant difference in between measurement methods (p = 0.339). Interobserver error analysis showed no statistical differences between the two observers. All measurements had excellent intraobserver reliability. ConclusionPreoperative assessment of glenoid version and inclination is crucial in ensuring precise implant positioning and orientation in TSA and RSA. This study observed a significant level of concordance between manual and automated measuring techniques utilizing MRI and CT scans. Mean glenoid inclination exhibited a statistically significant difference of less than 1° across the assessment modalities and no difference for glenoid version was noted. It seems to be questionable if this finding is clinically relevant. MRI may serve as a viable and safe option for assessing glenoid morphology, version and inclination if CT scans are not available.

lower lobe lung tumors move with amplitudes of up to 2 cm due to respiration. To reduce respiration imaging artifacts in planning CT scans, 4D imaging techniques are used. Currently, we use a single (midventilation) frame of the 4D data set for clinical delineation of structures and radiotherapy planning. A single frame, however, often contains artifacts due to breathing irregularities, and is noisier than a conventional CT scan since the exposure per frame is lower. Moreover, the tumor may be displaced from the mean tumor position due to hysteresis. The aim of this work is to develop a framework for the acquisition of a good quality scan representing all scanned anatomy in the mean position by averaging transformed (deformed) CT frames, i.e., canceling out motion. A nonrigid registration method is necessary since motion varies over the lung. 4D and inspiration breath-hold (BH) CT scans were acquired for 13 patients. An iterative multiscale motion estimation technique was applied to the 4D CT scan, similar to optical flow but using image phase (gray-value transitions from bright to dark and vice versa) instead. From the (4D) deformation vector field (DVF) derived, the local mean position in the respiratory cycle was computed and the 4D DVF was modified to deform all structures of the original 4D CT scan to this mean position. A 3D midposition (MidP) CT scan was then obtained by (arithmetic or median) averaging of the deformed 4D CT scan. Image registration accuracy, tumor shape deviation with respect to the BH CT scan, and noise were determined to evaluate the image fidelity of the MidP CT scan and the performance of the technique. Accuracy of the used deformable image registration method was comparable to established automated locally rigid registration and to manual landmark registration (average difference to both methods < 0.5 mm for all directions) for the tumor region. From visual assessment, the registration was good for the clearly visible features (e.g., tumor and diaphragm). The shape of the tumor, with respect to that of the BH CT scan, was better represented by the MidP reconstructions than any of the 4D CT frames (including MidV; reduction of "shape differences" was 66%). The MidP scans contained about one-third the noise of individual 4D CT scan frames. We implemented an accurate method to estimate the motion of structures in a 4D CT scan. Subsequently, a novel method to create a midposition CT scan (time-weighted average of the anatomy) for treatment planning with reduced noise and artifacts was introduced. Tumor shape and position in the MidP CT scan represents that of the BH CT scan better than MidV CT scan and, therefore, was found to be appropriate for treatment planning.

The purpose of CT simulation in radiotherapy is to acquire patient geometrical information and to build a patient geometrical model for treatment planning. Errors in patient model caused by motion artifacts will influence all treatment fractions and therefore should be handled carefully. Due to the tumor respiratory motion, the captured tumor position and shape can be heavily distorted. The distortions along the axis of motion could result in either a lengthening or shortening of the target. The center of the imaged target can be displaced by as much as the amplitude of the motion.A newly developed technique that can reduce motion artifacts and provide patient geometry throughout the whole breathing cycle is called respiration‐correlated or 4D CT scan. The basic idea for 4D CT scan is that, at every position of interest along patient's long axis, images are over‐sampled and each image is tagged with breathing phase information. After the scan is done, images are sorted based on the corresponding breathing phase signals. Thus, many 3D CT sets are obtained, each corresponding to a particular breathing phase, and together constitutes a 4D CT set that covers that the whole breathing cycle. 4D CT scan has been developed at various institutions with slightly different flavors. In this lecture, we will provide an overview of various implementations of 4D CT scan.4D CT scan can be used to account for respiratory motion to generate images with less distortion than 3D CT scan. 4D images also contain respiratory motion information of tumor and organs that is not available in a 3D CT image. This technology can be used for respiratory‐gated treatment to identify the patient‐specific phase of minimum tumor motion, determine residual tumor motion within the gate interval, and compare treatment plans at different phases. It can also be used for non‐gated treatment planning to define ITV by combining gross tumor volume at all breathing phases or using a method called maximum intensity projection. Of course 4D CT will also play a vital role in the futuristic 4D radiotherapy where the tumor is tracked dynamically during the treatment using multi‐leaf collimator.Existing problems for 4D CT scan include the increased imaging dose, CT tube heating, and data management. More importantly, one has to keep in mind that 4D CT scan is not really 4D. Temporal information is mapped into one breathing cycle. Irregular respiration will cause artifacts in 4D CT images. Patient coaching can improve the regularity of breathing pattern and thus reduce the residual artifacts. However this issue still deserves further studies.Educational Objectives:1. Understand the origin and magnitude of motion artifacts in free breathing helical CT scan.2. Understand how 4D CT scan works.3. Understand how 4D CT can be used in radiotherapy.4. Understand the remaining artifacts in 4D CT scan and possible future improvements.

The purpose of CT simulation in radiotherapy is to acquire patient geometrical information and to build a patient geometrical model for treatment planning. Errors in patient model caused by motion artifacts will influence all treatment fractions and therefore should be handled carefully. Due to the tumor respiratory motion, the captured tumor position and shape can be heavily distorted. The distortions along the axis of motion could result in either a lengthening or shortening of the target. The center of the imaged target can be displaced by as much as the amplitude of the motion. A newly developed technique that can reduce motion artifacts and provide patient geometry throughout the whole breathing cycle is called respiration‐correlated or 4D CT scan. The basic idea for 4D CT scan is that, at every position of interest along patient's long axis, images are over‐sampled and each image is tagged with breathing phase information. After the scan is done, images are sorted based on the corresponding breathing phase signals. Thus, many 3D CT sets are obtained, each corresponding to a particular breathing phase, and together constitutes a 4D CT set that covers that the whole breathing cycle. 4D CT scan has been developed at various institutions with slightly different flavors. In this lecture, we will provide an overview of various implementations of 4D CT scan. 4D CT scan can be used to account for respiratory motion to generate images with less distortion than 3D CT scan. 4D images also contain respiratory motion information of tumor and organs that is not available in a 3D CTimage. This technology can be used for respiratory‐gated treatment to identify the patient‐specific phase of minimum tumor motion, determine residual tumor motion within the gate interval, and compare treatment plans at different phases. It can also be used for non‐gated treatment planning to define ITV by combining gross tumor volume at all breathing phases or using a method called maximum intensity projection. Of course 4D CT will also play a vital role in the futuristic 4D radiotherapy where the tumor is tracked dynamically during the treatment using multi‐leaf collimator. Existing problems for 4D CT scan include the increased imagingdose,CT tube heating, and data management. More importantly, one has to keep in mind that 4D CT scan is not really 4D. Temporal information is mapped into one breathing cycle. Irregular respiration will cause artifacts in 4D CTimages. Patient coaching can improve the regularity of breathing pattern and thus reduce the residual artifacts. However this issue still deserves further studies. Educational Objectives: 1. Understand the origin and magnitude of motion artifacts in free breathing helical CT scan. 2. Understand how 4D CT scan works. 3. Understand how 4D CT can be used in radiotherapy. 4. Understand the remaining artifacts in 4D CT scan and possible future improvements.

Introduction Odontoid fractures (OF) are frequent in the trauma population, and there is no universally accepted single method of management. The objective of this study was to evaluate the epidemiology and management of type II OF in Latin America treated either with rigid cervical orthosis or surgery. Patients and Methods A total of 83 patients treated conservatively or by surgery were enrolled in this retrospective study. Medical charts, imaging studies, and outcomes of patients were analyzed in the pretreatment period and at the last medical evaluation. The fractures were assessed using conventional radiographs, three-dimensional computed tomographic (3D-CT) scans, and magnetic resonance images. Fracture gaps, the direction and the degree of displacement of the odontoid process, the fracture line anatomy, the degrees of atlantoaxial instability, the comminuted fracture, and the surface contact area were analyzed. The decision for operative or nonoperative treatment was based on anesthesia risk, and patient's choice of the nonoperative treatment. The nonoperative management generally consisted of a rigid cervical orthosis for 3 months. The type of surgery to be performed was chosen by the surgeon. The solid bony union was defined as the presence of bony bridges and the definite continuity of cortical bone. Fibrous union was considered present when no degree of motion was evident in dynamic radiographs despite persistent cortical bone discontinuity within a fracture gap on 3D CT scans. Nonunion was defined as a definite fracture gap with abnormal motion of the fractured dens on dynamic radiographs and on a 3D CT scan. Results A total of 83 patients were included in this study. The patients were 78.3% men, the mean age = 44.98 ( ± 23.20 years) years. Traffic accidents (66.3%) were the most common cause of trauma. The main symptom was pain (85.5%) in the posterior cervical region. The median time elapsed from accident to surgery was 7 days (P25: 2/P75: 27.5). Median follow-up was 23.66 ( ± 25.43 months) months. Conservative treatment with cervical orthosis, for example, Miami J collar or halo-vest was used in 20.5% of the cases. Odontoid screw technics (57.6%) were the most common surgical treatment adopted as primary surgical treatment. Symptomatic nonunion was observed in two cases with conservative treatment and three cases after odontoid screw fixation. All the patients were referred to posterior C1–C2 fixation. The posterior fixation tended to be used after conservative failed therapy, after nonunion anterior screw surgery, and in fractures with greater displacement. The most common radiological feature was no displacement of the odontoid process in relation to the body of C2, horizontal fracture line, gap fracture &lt; 2 mm, no subluxation across each C1–C2 facet joint and no comminuted fracture. Conclusion The patients treated nonoperatively with a rigid collar may have an overall favorable outcome compared with surgical treatment. A well-designed prospective study is needed to better elucidate optimal treatment algorithms from both an outcomes and cost-effectiveness perspective.

3D CT scan is actually the gold standard for preoperative diagnosis of pelvic discontinuity (PD) in hip revision surgery. Aim of this study was to compare the accuracy of 3D-modeling with traditional and 3D CT scan. We retrospectively identified 56 patients who underwent total hip arthroplasty revisions with Paprosky Type-3 periacetabular bone defects. Preoperative X-rays, CT scans and 3D-models were blindly reviewed by two orthopedic surgeons to detect possible pelvic discontinuities. Results were compared with surgical notes. Independent sensitivities, specificities, positive predictive values and negative predictive values were calculated for X-rays, CT scan and 3D models. Analysis of interobserver reliability was performed. Fifty-six patients met inclusion criteria. In nine patients, surgical notes indicated a pelvic discontinuity. On 3D CT scans, PD was identified in 25 cases for observer 1 and in 24 cases for observer 2. Analyzing 3D-models, PD was identified in eleven patients by both observers. The nine patients, with PD reported on the surgical report, were all identified with both the techniques. The specificity of standard 3D CT was 0.66 for observer 1 and 0.68 for observer 2 and increased to 0.96 for both observers with the utilization of 3D-models. The positive predictive value increased from 0.36 (observer 1) and 0.38 (observer 2) with the CT evaluation to 0.82 in the 3D-models evaluation. The analysis of 3D models was characterized by a perfect intraobserver reliability (intraobserver correlation coefficient = 1). The observers showed substantial agreement for PD classification; the kappa values were 0.96 and 0.77, respectively, for CT scan and 3D-model evaluation. 3D-modeling showed higher specificity than traditional and 3D CT scans in identification of PD in Paprosky Type-3 periacetabular bone defects.

The Wrightington classification system of fracture-dislocations of the elbow divides these injuries into six subtypes depending on the involvement of the coronoid and the radial head. The aim of this study was to assess the reliability and reproducibility of this classification system. This was a blinded study using radiographs and CT scans of 48 consecutive patients managed according to the Wrightington classification system between 2010 and 2018. Four trauma and orthopaedic consultants, two post CCT fellows, and one speciality registrar based in the UK classified the injuries. The seven observers reviewed preoperative radiographs and CT scans twice, with a minimum four-week interval. Radiographs and CT scans were reviewed separately. Inter- and intraobserver reliability were calculated using Fleiss and Cohen kappa coefficients. The Landis and Koch criteria were used to interpret the strength of the kappa values. Validity was assessed by calculating the percentage agreement against intraoperative findings. Of the 48 patients, three (6%) had type A injury, 11 (23%) type B, 16 (33%) type B+, 16 (33%) Type C, two (4%) type D+, and none had a type D injury. All 48 patients had anteroposterior (AP) and lateral radiographs, 44 had 2D CT scans, and 39 had 3D reconstructions. The interobserver reliability kappa value was 0.52 for radiographs, 0.71 for 2D CT scans, and 0.73 for a combination of 2D and 3D reconstruction CT scans. The median intraobserver reliability was 0.75 (interquartile range (IQR) 0.62 to 0.79) for radiographs, 0.77 (IQR 0.73 to 0.94) for 2D CT scans, and 0.89 (IQR 0.77 to 0.93) for the combination of 2D and 3D reconstruction. Validity analysis showed that accuracy significantly improved when using CT scans (p = 0.018 and p = 0.028 respectively). The Wrightington classification system is a reliable and valid method of classifying fracture-dislocations of the elbow. CT scans are significantly more accurate than radiographs when identifying the pattern of injury, with good intra- and interobserver reproducibility. Cite this article: Bone Joint J 2020;102-B(8):1041-1047.

In many patients respiratory motion causes motion artifacts in CT images, thereby inhibiting precise treatment planning and lowering the ability to target radiation to tumors. The 4D Phantom, which includes a 3D stage and a 1D stage that each are capable of arbitrary motion and timing, was developed to serve as an end-to-end radiation therapy QA device that could be used throughout CT imaging, radiation therapy treatment planning, and radiation therapy delivery. The dynamic accuracy of the system was measured with a camera system. The positional error was found to be equally likely to occur in the positive and negative directions for each axis, and the stage was within 0.1 mm of the desired position 85% of the time. In an experiment designed to use the 4D Phantom's encoders to measure trial-to-trial precision of the system, the 4D Phantom reproduced the motion during variable bag ventilation of a transponder that had been bronchoscopically implanted in a canine lung. In this case, the encoder readout indicated that the stage was within 10 microns of the sent position 94% of the time and that the RMS error was 7 microns. Motion artifacts were clearly visible in 3D and respiratory-correlated (4D) CT scans of phantoms reproducing tissue motion. In 4D CT scans, apparent volume was found to be directly correlated to instantaneous velocity. The system is capable of reproducing individual patient-specific tissue trajectories with a high degree of accuracy and precision and will be useful for end-to-end radiation therapy QA.

Four dimensional computed tomography (4DCT) scans are routinely used in radiation therapy to determine the internal treatment volume for targets that are moving (e.g. lung tumors). The use of these studies has allowed clinicians to create target volumes based upon the motion of the tumor during the imaging study. The purpose of this work is to determine if a target volume based on a single 4DCT scan at simulation is sufficient to capture thoracic motion. Phantom studies were performed to determine expected differences between volumes contoured on 4DCT scans and those on the evaluation CT scans (slow scans). Evaluation CT scans acquired during treatment of 11 patients were compared to the 4DCT scans used for treatment planning. The images were assessed to determine if the target remained within the target volume determined during the first 4DCT scan. A total of 55 slow scans were compared to the 11 planning 4DCT scans. Small differences were observed in phantom between the 4DCT volumes and the slow scan volumes, with a maximum of 2.9%, that can be attributed to minor differences in contouring and the ability of the 4DCT scan to adequately capture motion at the apex and base of the motion trajectory. Larger differences were observed in the patients studied, up to a maximum volume difference of 33.4%. These results demonstrate that a single 4DCT scan is not adequate to capture all thoracic motion throughout treatment.

Introduction:Reliability is the study of internal consistency, reproducibility (intraobserver and interobserver), and agreement. Reproducibility studies that classify tibial plateau fractures have used plain radiography and two-dimensional (2D) CT scans and three-dimensional (3D) printing. The objective of this study was to evaluate the reproducibility of the Luo Classification of tibial plateau factures and the surgical approaches chosen for these fractures based on 2D CT scans and 3D printing.Methods:A reliability study was performed at the Universidad Industrial de Santander, Colombia, that evaluated the reproducibility of the Luo Classification of tibial plateau fractures and the choice of surgical approaches based on 20 CT scans and 3D printing, with five evaluators.Results:For the trauma surgeon, reproducibility was better when evaluating the classification using 3D printing, with a kappa of 0.81 (95% confidence interval [CI], 0.75-0.93; P < 0.01) than when using CT scans, with a kappa of 0.76 (95% CI, 0.62-0.82; P < 0.01). When comparing the surgical decisions made by the fourth-year resident with those of the trauma surgeon, a fair reproducibility was obtained using CT, with a kappa of 0.34 (95% CI, 0.21-0.46; P < 0.01), which improved to substantial when using 3D printing, with a kappa of 0.63 (95% CI, 0.53-0.73; P < 0.01).Discussion:This study found that 3D printing provided more information than CT and decreased measurement errors, thereby improving reproducibility, as shown by the higher kappa values that were obtained.Conclusion:The use of 3D printing and its usefulness are helpful to decision making when providing emergency trauma services to patients with intraarticular fractures such as those of the tibial plateau.

e13530 Background: To explore and compare the dosimetric variance in forward intensity modulated radiotherapy (IMRT) based on 4D CT and 3D CT after breast conserving surgery. Methods: Seventeen patients after breast conserving surgery underwent the 3D CT simulation scans followed by respiration-synchronized 4D CT simulation scans on the state of free breathing. The treatment plan constructed using the end inspiration (EI) scan was then copied and applied to the end expiration (EE) and 3D scans and the dose distribution was calculated separately. Dose–volume histograms (DVHs) parameters for the CTV, PTV, ipsilateral lung (IPSL) and heart were evaluated and compared. Results: The CTV volume amplitude was 11.93 ± 28.64 cm3, and volume of the CTV receiving 95%, 100%, and 103% prescription dose among different scans were all differed by &lt; 0.4%. Mean PTV dose at EE was lower than EI (t = 2.87, p = 0.011), but there were no statistice significance between 3D CT scan and EI, EE scans (t = 1.06, -1.59; p = 0.304, 0.132). The homogeneity index (HI) at EI, EE, 3D plans were 0.156 ± 0.02, 0.162 ± 0.02, 0.161 ± 0.02, respectively, and difference only between EI and EE (t = -2.56, p = 0.021). The highest conformal index (CI) was at EI phase (t = 4.55, 2.70; p = 0.000, 0.016), and there was no significant difference between EE and 3D (t = 0.04, p = 0.967). The V20, V30, V40, V50 and Dmean of IPSL at EE phase were lower than EI (t = 2.39~5.54, p = 0.000~0.030). There were no significant differences in all the indexes for heart (t = -1.77~1.40, p = 0.128~0.693). Conclusions: The breast deformation during respiration may be disregarded in whole breast IMRT; PTV dose distribution was changed significantly between EI and EE phase, and the differentiation of the lung high dose area between EI and EE phase may induced by thorax expansion. 3D treatment planning is sufficient for whole breast forward IMRT, but 4D CT scans assist with respiratory gating ensure precise delivery of radiation dose.

Separation of left and right lungs from binary segmentation is often necessary for quantitative image-based pulmonary disease evaluation. In this article, we present a new fully automated approach for accurate, robust, and efficient lung separation using 3-D CT scans. Our method follows a hysteresis setting that utilizes information from both lung regions and background gaps. First, original segmentation is separated by subtracting the gaps between left and right lungs, which are enhanced with Hessian filtering. Second, the 2-D separation manifold in 3-D image space is estimated based on the distance information from the two subsets. Finally, the separation manifold is projected back to the original segmentation in order to produce the separated lungs through optimization for addressing minor local variations. An evaluation on over 400 human and 100 small animal 3-D CT images with various abnormalities is performed. The proposed scheme successfully separated all connections on the candidate CT images. Using hysteresis mechanism, each phase is performed robustly and 3-D information is utilized to achieve a generic, efficient, and accurate solution.

The question remains regarding the dosimetric impact of intrafraction motion in 3D breast treatment. This study was conducted to investigate this issue utilizing the 4DCT scan. The 4D and helical CT scan sets were acquired for 12 breast cancer patients. For each of these patients, based on the helical CT scan, a conventional 3D conformal plan was generated. The breast treatment was then simulated based on the 4DCT scan. In each phase of the 4DCT scan, dose distribution was generated with the same beam parameters as the conventional plan. A software package was developed to compute the cumulative dose distribution from all the phases. Since the intrafraction organ motion is reflected by the 4DCT images, the cumulative dose computed based on the 4DCT images should be closer to what the patient received during treatment. Various dosimetric parameters were obtained from the plan and 4D cumulative dose distribution for the target volume and heart, and were compared to deduce the motion-induced impacts. The studies were performed for both whole breast and partial breast treatment. In the whole breast treatment, the average intrafraction motion induced changes in D95, D90, V100, V95, and V90 of the target volume were -5.4%, -3.1%, -13.4%, -5.1%, and -3.2%, respectively, with the largest values at -26.2%, -14.1%, -91.0%, -15.1%, and -9.0%, respectively. Motion had little impact on the Dmax of the target volume, but its impact on the Dmin of the target volume was significant. For left breast treatment, the motion-induced Dmax change to the heart could be negative or positive, with the largest increase at about 6 Gy. In partial breast treatment, the only non-insignificant impact was in the Dmin of the CTV (ranging from -15.2% to 11.7%). The results showed that the intrafraction motion may compromise target dose coverage in breast treatments and the degree of that compromise was correlated with motion magnitude. However, the dosimetric impact of the motion on the heart dose may be limited.

Revision total hip arthroplasty (rTHA) presents significant challenges, particularly in patients with severe acetabular bone defects. Traditional treatment options often fall short, leading to the emergence of custom-made 3D-printed acetabular implants. Accurate assessment of implant positioning is crucial for ensuring optimal postoperative outcomes and for providing feedback to the surgical team. This single-center, retrospective cohort study evaluates the accuracy of standard 2D radiographs versus 3D CT scans in assessing the positioning of these implants, aiming to determine if 2D imaging could serve as a viable alternative for the postoperative evaluation. We analyzed the implant positions of seven rTHA patients with severe acetabular defects (Paprosky ≥ Type IIIA) using an alignment technique that integrates postoperative 2D radiographs with preoperative CT plans. Two independent investigators, one inexperienced and one experienced, measured the positioning accuracy with both imaging modalities. Measurements included translational shifts from the preoperatively templated implant position in the craniocaudal (CC), lateromedial (LM), and ventrodorsal (VD) directions, as well as rotational differences in anteversion (AV) and inclination (INCL). The study demonstrated that 2D radiographs, when aligned with preoperative CT data, could accurately assess implant positions with precision nearly comparable to that of 3D CT scans. Observed deviations were 1.4 mm and 2.7 mm in CC and LM directions, respectively, and 3.6° in AV and 0.7° in INCL using 2D imaging, all within clinically acceptable ranges. For 3D CT assessments, mean interobserver variability was up to 0.9 mm for translational shifts and 1.4° for rotation, while for 2D alignment, observer differences were 1.4 mm and 3.2° for translation and rotation, respectively. Comparative analysis of mean results from both investigators, across all dimensions (CC, LM, AV, and INCL) for 2D and 3D matching, showed no significant differences. In conclusion, conventional anteroposterior 2D radiographs of the pelvis can sufficiently determine the positioning of custom-made acetabular implants in rTHA. This suggests that 2D radiography is a viable alternative to 3D CT scans, potentially enhancing the implementation and quality control of advanced implant technologies.