ORIGINAL ARTICLE Multi-center study of inter-rater reproducibility, image quality, and diagnostic accuracy of CZT versus conventional SPECT myocardial perfusion imaging Nabeel Mansour, MD,a,b Stephan G. Nekolla, PhD,a,c Eliana Reyes, MD, PhD,d George Angelidis, MD, PhD,e Panagiotis Georgoulias, MD, PhD,e Constantinos Anagnostopoulos, MD, PhD, FESC, FRCR, FRCP,f Paco Bravo, MD,g Isabella Bruno, MD,h Albert Flotats, MD,i Francisco Fuentes-Ocampo, MD,i Roberto Sciagrà, MD,j Felix Keng, MD,k Lukas Kessler, MD,n Maria Papathanasiou, MD,l Prem Soman, MD, PhD, FACC, FRCP,m and Christoph Rischpler, MDn a Department of Nuclear Medicine, Klinikum Rechts der Isar, Technical University Munich, Munich, Germany b Department of Radiology, University Hospital, LMU Munich, Munich, Germany c DZHK (German Centre for Cardiovascular Research), Partner Site Munich Heart Alliance, Munich, Germany d Royal Brompton and Harefield NHS Trust, London, UK e Department of Nuclear Medicine, Faculty of Medicine, University of Thessaly, University Hospital of Larissa, Larissa, Greece f BRFAA – Biomedical Research Foundation Academy of Athens, Athens, Greece g Divisions of Nuclear Medicine and Cardiology, Departments of Radiology and Medicine, Hospital of the University of Pennsylvania, Philadelphia, PA h Fondazione Policlinico Universitario Agostino Gemelli IRCCS, Roma, Italy i Nuclear Medicine Department, Hospital de la Santa Creu i Sant Pau, Universitat Autònoma de Barcelona, Barcelona, Spain j Department of Experimental and Clinical Biomedical Sciences ‘Mario Serio’, University of Florence, Florence, Italy k National Heart Centre, Singapore, Singapore l Department of Cardiology, West German Heart and Vascular Center, University Hospital, Essen, Germany m Division of Cardiology, University of Pittsburgh Medical Center, Pittsburgh, PA n Department of Nuclear Medicine, University Hospital Essen, University of Duisburg-Essen, Essen, Germany Received Apr 1, 2022; accepted Jun 16, 2022 doi:10.1007/s12350-022-03054-w Supplementary Information The online version contains supple- Nabeel Mansour and Stephan G. Nekolla have contributed equally to mentary material available at https://doi.org/10.1007/s12350-022- this work. 03054-w. Reprint requests: Nabeel Mansour, MD, Department of Nuclear The authors of this article have provided a PowerPoint file, available Medicine, Klinikum Rechts der Isar, Technical University Munich, for download at SpringerLink, which summarises the contents of the Munich, Germany; Nabeel.Mansour@med.uni-muenchen.de paper and is free for re-use at meetings and presentations. Search for J Nucl Cardiol 2023;30:528–39. the article DOI on SpringerLink.com. 1071-3581/$34.00 The authors have also provided an audio summary of the article, which Copyright  2022 The Author(s) is available to download as ESM, or to listen to via the JNC/ASNC Podcast. 528 Journal of Nuclear Cardiology Mansour et al 529 Volume 30, Number 2;528–39 Multi-center study of inter-rater reproducibility, image quality… Background. Cadmium-zinc-telluride (CZT)-based detectors exhibit higher diagnostic sensitivity in myocardial perfusion imaging (MPI) than conventional Anger-MPI for detection of coronary artery disease (CAD); however, reduced specificity and diagnostic accuracy of CZT-MPI were observed. This study aims to compare these different camera systems and to examine the degree of inter-rater reproducibility among readers with varying experience in MPI. Methods. 83 patients who underwent double stress/rest examinations using both a CZT and conventional SPECT cameras within one visit were included. Anonymized and randomized MPI-images were distributed to 15 international readers using a standardized questionnaire. Subsequent coronary angiography findings of ten patients served as a reference for analysis of sensitivity and specificity. Results. Image quality was significantly better in CZT-MPI with significantly lower breast attenuation (P < 0.05). CZT-MPI exhibited higher sensitivity than Anger-MPI (87.5% vs. 62.5%) and significantly reduced specificity (40% vs. 100%). Readers experienced with both camera systems had the highest inter-rater agreement indicating higher reproducibility (CZT 0.54 vs. conv. 0.49, P < 0.05). Conclusions. Higher diagnostic sensitivity of CZT-MPI offers advantages in detection of CAD yet potentially of at the cost of reduced specificity, therefore it requires special training and a differentiated evaluation approach, especially for non-experienced readers with such camera systems. (J Nucl Cardiol 2023;30:528–39.) Key Words: cadmium-zinc-telluride Æ myocardial perfusion imaging Æ coronary artery disease Æ single-photon emission tomography iodide (NaI (Tl)) detectors used in conventional Anger Abbreviations cameras.7 The CZT-based camera used in this study (D- MPI Myocardial perfusion imaging SPECT; Spectrum Dynamics, Israel) is equipped with SPECT Single photon emission computed tungsten parallel-hole collimators, which show a signif- tomography icant improvement in geometric efficiency compared to CZT Cadmium-zinc-telluride standard lead parallel-hole low-energy, high-resolution CAD Coronary artery disease (LEHR) collimators used in Anger cameras.5 Addition- AHA American Heart Association ally, the open design of the CZT camera and the Tc-99m Technetium-99m flexibility of the L-shaped detector columns enable MIBI Sestamibi better accessibility and acquisition in a semi-supine MBq Megabecquerel position which serves as an advantage for large sized AC Attenuation correction patients or patients with orthopnea. However, this comes EF Ejection fraction potentially at the cost of reduced specificity in obese patients examined in this position, as argued in multiple studies.8,9 INTRODUCTION In contrary to Anger-MPI, the ability to image upright and therefore the tendency for imaging of a Myocardial perfusion imaging (MPI) is one of the mainly obese patients may result in reduced diagnostic most often used techniques worldwide for the detection specificity with inferolateral segments still suffering and assessment of coronary artery disease (CAD). The marked soft tissue attenuation on CZT cameras.9 Fur- diagnostic performance of conventional single-photon thermore, most dedicated CZT cameras for MPI are not emission computed tomography (Anger-SPECT) is well equipped with an integrated computed tomography (CT) established, with current data yielding estimates for component with most studies performed without atten- sensitivity ranging between 0.82 and 0.91, and for uation correction (AC), harboring the risk of attenuation specificity between 0.70 to 0.90, depending on clinical artifacts being interpreted as perfusion abnormalities.10 history and the cutoff chosen for significant CAD.1 In With the introduction of high speed solid-state the last decade, rapid-acquisition dedicated cardiac cardiac cameras first studies comparing these technolo- camera systems with cadmium-zinc-telluride (CZT)- gies suggested high correlation in image quality based detectors have been introduced into clinical compared to conventional SPECT, with an equivalent practice.2–6 CZT-detectors were selected because of level of diagnostic confidence.11–13 In a meta-analysis their superior energy resolution compared to sodium performed by Nudi et al.8, CZT-MPI presented a 530 Mansour et al Journal of Nuclear Cardiology Multi-center study of inter-rater reproducibility, image quality… March/April 2023 relatively high rate of false positive results and reduced than 90 days after MPI, were available in ten patients. specificity particularly in imaging of obese patients, As part of the clinical routine, cardiovascular risk factors respectively. In the majority of the studies included, (arterial hypertension, diabetes mellitus, dyslipidemia, patient cohorts were paired based on specific clinical nicotine abuse, and history of nicotine abuse) were characteristics. present in 64 patients. The objective of this study was to examine inter- rater reproducibility, diagnostic certainty, reader expe- Readers rience effects between CZT-MPI and conventional Anger-MPI in the same patient cohort in a multi-center 15 international readers from the field of nuclear analysis with readers of varying experience in the cardiology with varying experience in the interpretation assessment of SPECT-MPI. In addition, we delineated of SPECT-MPI were recruited. They were divided based whether and to what extent the two camera systems, on underlying experience into three groups of five differ in terms of image quality and susceptibility to participants each: Five readers with experience with artifacts. both conventional SPECT cameras and a CZT-SPECT camera (group 1); five readers who only have experience with conventional SPECT cameras (group 2) and five METHODS readers with little or no experience in MPI interpretation (group 3). The anonymized and randomized MPI images Study design were distributed in the form of static images of the short Retrospective multicentric reading study with 15 axes and horizontal/vertical long axes as well as the international readers from the field of nuclear cardiology polar maps by means of an online study. The online with varying experience in MPI. The readers evaluated study was carried out using Evasys Survey Grid15 with anonymized and randomized MPI findings of 83 patients an encrypted invitation letter containing a Uniform acquired with a CZT-based as well as with one of two Resource Locator (URL) address with an individual conventional Anger SPECT cameras (total of 166 transaction number (TAN) password for each reader. A cases). Informed consent was waived due to the retro- standardized questionnaire broadly used in the interpre- spective character of the study. tation of MPI with questions regarding image quality, artifacts, final diagnosis, diagnostic certainty and the 17- segment model of the American Heart Association Patients (AHA)16 was used. This standardized questionnaire is Of the 361 patients who underwent myocardial used for purely image-based evaluation of data, which perfusion imaging (MPI) in the year 2013 as part of the make up the main building blocks of MPI findings. No clinical routine in the Klinikum rechts der Isar, Munich, clinical information was provided, and all readers were Germany, we identified 83 patients who were sequen- blinded to the modality of examination (CZT- or tially examined with a CZT (D-SPECT, Spectrum conventional SPECT-MPI). Dynamics) and with one of the two available conven- tional SPECT cameras (Siemens E.CAM or Symbia T6, SPECT cameras Siemens Healthineers) within one visit both after stress and at rest. Data were originally collected in routine The recorded light pulses are amplified with pho- clinical practice at our institute when the CZT camera toelectron multipliers (PMT) that deliver electrical was newly introduced and were used for comparison pulses proportional to photon energy. In this way, the with the conventional SPECTs to familiarize the readers activity of the radiotracer in the examined organ is with this new technology. The pretest probability for the visualized as a projection onto a plane. The gamma presence or absence of significant obstructive CAD, camera, which rotates around the patient and depicts defined as a C 70% stenosis of at least one coronary isotope activity in multiple angle projections, enables artery was calculated using a consortium score to the three-dimensional reconstruction of isotope activity validate the indication for non-invasive diagnostics.14 and distribution in the body. Two conventional cameras In a proportion of patients, prior coronary angiographic were used in this study (Siemens E.CAM and Symbia T, data indicated the extent of vessel involvement Siemens Medical Solutions). A D-SPECT system (Spec- (N = 41). Other patients received a subsequent coronary trum Dynamics, Israel) was used for CZT-MPI.17 This angiography depending on symptoms and the results of system uses pixilated CZT crystal detector columns the non-invasive diagnostic work-up. Relevant coronary mounted in nine vertical columns that are arranged in a angiography results, defined as results acquired no later 90  geometry. No attenuation correction (AC) was Journal of Nuclear Cardiology Mansour et al 531 Volume 30, Number 2;528–39 Multi-center study of inter-rater reproducibility, image quality… applied in this study, since the D-SPECT camera used at Statistical analysis the time of the study did not support it. Statistical analysis was performed using R statisti- cal software (R version 3.6.3). The Wilcoxon test was SPECT protocols used to test differences between the two camera tech- nologies or between two groups of readers.21 When 79 patients were examined via one-day protocol and individual values occurred very frequently, the sign test four patients via two-day protocol. The patients with the was calculated.22 Statistical significance was denoted by one-day protocol were examined sequentially with both P values \ .05. Inter-rater agreement was analyzed with camera technologies, first in stress and then in rest (66 the Fleiss’ Kappa statistics.23 The analysis of the Inter- times CZT-stress-MPI first and 13 times conventional rater agreement was divided in two categories, the first stress MPI first). Mean time interval between the stress regarding all answers for diagnosis (normal, ischemia, examinations was approximately 33 (± 31) minutes mixed ischemia/scar and scar), and the second regarding when CZT-MPI was performed as first procedure vs. 58 abnormality in comparison of normal versus abnormal (± 15) minutes when Anger-MPI was performed as first findings (normal versus all other answers). Sensitivity, procedure. specificity, and accuracy were calculated using the Injected activities of the perfusion tracer Tc-99m angiography group (N = 10) and a negative control sestamibi were body-weight adjusted according to cur- group which did not undergo angiography after non- rent guidelines18 resulting in mean dose of 414 (± 221) invasive exclusion of CAD (N = 8). A positive finding MBq for the stress study and 838 (± 282) MBq for the was defined at a cutoff of C 70% luminal diameter rest study in the one-day protocol. Using the two-day- narrowing of at least one coronary vessel. Eight patients protocol injected activities were 518 (± 255) MBq of the angiography group displayed coronary artery (stress) and 548 (± 230) MBq (rest), respectively. On stenosis with the cutoff as defined above and two conventional SPECT cameras, all patients were imaged patients had normal angiograms. A total of ten patients in a supine position and on the CZT-SPECT system all were defined as a negative group consisting of patients patients were imaged using the standard semi-supine with no relevant coronary artery stenosis in coronary position. angiography (N = 2) as well as patients who had Scan durations were 20 min fixed per scan in the intermediate pretest probability for CAD, yet no indi- conventional cameras and on average 12 (stress) resp. 5 cation for a coronary angiography after revision of all (rest) minutes the D-SPECT using count limits. non invasive examinations and available data and exclusion of CAD in interdisciplinary consensus Image reconstruction and interpretation (N = 8). The individual coronary territories were compared Analysis of the volumetric data on the regional in accordance with the 17-segment model of the AHA. tracer distribution resulting from MPI was performed A positive MPI finding was defined as an SDS of at least with a dedicated software (MunichHeart).19 This quan- 2 and/or an SSS of at least 4. The SRS was assessed titative analysis software enables the quantification and analogously to the SSS. thus the objectification of tracer accumulation in the heart both at rest and under stress. It has been validated in multiple phantom studies and in several data sets of RESULTS healthy patients and patients with CAD of various degrees of severity.20 After volumetric data analysis and Study population image reconstruction, radiotracer activity is assessed in 83 patients aged between 38 and 91 years (mean, sections along the short and long axes of the left 67.7 ± 11.6) were enrolled. 49 patients (59%) were ventricle. Furthermore, a polar map of the radiotracer male. The body mass index (BMI) was between 16.5 and distribution throughout the left ventricle is generated. 55.4 kg/m2. Cardiovascular risk factors were docu- A standardized myocardial scoring system for the mented only in 64 patients and 41 patients had pre- resulting polar maps is available for quantification of existing CAD as summarized in Table 1. Ten patients results (summed stress score (SSS), summed rest score had coronary angiography results relevant to this study (SRS), and summed difference score (SDS) as described defined as results acquired within a maximum period of by Berman et al.16. As displayed in Figure 1, different 90 days after MPI (median 14, range of 4 to 85 days). segments represent different areas covered by each of Coronary angiography revealed significant coronary the main coronary artery branches. Slight variations in artery stenosis, defined as C70% stenosis of vessel allocation of the segments depending on coronary artery lumen in eight patients, of whom 7 received dominance can be observed. 532 Mansour et al Journal of Nuclear Cardiology Multi-center study of inter-rater reproducibility, image quality… March/April 2023 Figure 1. Myocardial perfusion imaging acquired with a D-SPECT camera. A Polarmaps of the stress- (left) and rest examination (right). B Static images of the short axes and horizontal/vertical long axes. C: Schematic illustration of the 17-segment model of the left ventricle with illustration of supply areas of the coronary arteries. LAD left anterior descending artery, RCA right coronary artery, LCX ramus circumflexus. percutaneous transluminal coronary angioplasty (PTCA) (1.33 ± 0.18 for males vs. 1.48 ± 0.33 for females), yet and stenting. In one patient, a C 70% stenosis was not this difference was not significant (P=0.052). A signif- revascularized due to an unsuccessful recanalization icant difference in breast attenuation between female attempt of the occluded vessel. CAD was excluded in and male patients was observed only in MPI findings of two patients with normal angiograms. the Anger cameras and only by readers who only had experience with conventional SPECT cameras (Group 2, 1.3 ± 0.21 vs. 1.63 ± 0.45, P \ 0.05). Another possible Image quality and artifacts source of artifacts is liver and bowel activity. The CZT Image quality of the D-SPECT was rated signifi- camera exhibited slightly higher liver and bowel activ- cantly better compared to conventional SPECT cameras ity, but no significant difference for any reader group (2.8 ± 0.36 vs. 2.6 ± 0.33, P \ 0.05, Figure 2). In was observed (1.85 ± 0.65 vs. 1.71 ± 0.59, P= 0.073, addition, significantly less breast attenuation was Figure 3). When comparing liver and bowel activity in observed in the scans acquired with the CZT camera correlation with the time interval between the first and (1.2 ± 0.14 vs. 1.4 ± 0.26, P\ 0.05), regardless of reader second stress scan, a significant negative correlation was experience (Figure 3). When examining the relationship observed (ps \ 0.05). This observation was made by the between breast attenuation and gender (N = 49 male, N experienced readers (groups 1 and 2), but not by the = 34 female), the overall differences for CZT-MPI were inexperienced readers (group 3, P =0.074). Although the not significant (1.19 ± 0.14 for males vs. 1.17 ± 0.14 for stress MPI was performed as the first procedure in 80% females, P = 0.6). Higher breast attenuation was of the patients on the CZT camera, and a significant observed in females when examined with Anger-MPI negative correlation between time and liver and bowel Journal of Nuclear Cardiology Mansour et al 533 Volume 30, Number 2;528–39 Multi-center study of inter-rater reproducibility, image quality… Table 1. Patient characteristics and cardiological history Mean age (years) 67.7 ± 11.6 Male 49 (59%) Female 34 (41%) Cardiovascular risk factors (N=64) Hypertension 57 (89%) Diabetes mellitus 18 (28%) Dyslipidemia 43 (67%) Smoking 12 (19%) History of smoking 17 (27%) Not documented 19 (23%) Induced myocardial stress Exercise 20 (24%) Adenosine 52 (63%) Dobutamine 3 (4%) Regadenoson 3 (4%) Not documented 5 (6%) Cardiological history (N=41) Number of diseased vessels 0-vessel disease 20 1-vessel disease 4 2-vessel disease 5 3-vessel disease 12 Mean Agatston score 334.2 History of myocardial infarction 19 (46%) Mean SD Median Min Max Tc-99m-MIBI Dose [MBq] Rest 838.0 282.2 881.0 199 1404 Stress 414.4 221.0 337.5 180 1108 EF [%] Rest 60.0 11.6 60.0 23 80 Stress 59.2 12.1 59.0 24 87 Cardiovascular risk factors were documented in 64 patients. 41 patients had a positive cardiological history regarding CAD Tc-99m technetium-99m, MIBI sestamibi, MBq megabecquerel, EF ejection fraction, SD standard deviation artifacts was confirmed, this potential artifact was not different (0.39 ± 0.50 vs. 0.38 ± 0.53, P = 0.70, significantly higher for the CZT camera compared to Figure 4). Readers experienced with both camera conventional SPECT (1.85 ± 0.65 vs. 1.71 ± 0.59, systems (group 1) had significantly lower SSS and P = 0.073, Figure 3). In addition, visual assessment of SDS than the other groups, regardless of the camera the CZT camera images showed a higher signal-to-noise system used (SSS: 0.42 ± 0.63 CZT vs. 0.28 ± 0.55 ratio than that of the conventional cameras. conventional, P \ 0.05, SDS: 0.38 ± 0.63 CZT vs. 0.12 ± 0.27 conventional, P \ 0.05). In contrary, inexperi- enced readers (group 3) reached the highest SSS (0.97 ± Perfusion abnormalities 0.73 CZT vs. 0.77 ± 0.73 conventional, P = 0.0071*, Regarding the final diagnosis, all reader groups P \ 0.05). observed a significantly higher rate of abnormalities in CZT-MPI (0.5 ± 0.32 vs. 0.37 ± 0.30, P \ 0.05). This Inter-rater agreement and interpretative correlation was also apparent with significantly higher certainty mean SSS and SDS scores in all groups on CZT-MPI (SSS: 0.73 ± 0.63 vs. 0.54 ± 0.60, P = 0. 0.00028*, P Readers with experience with both camera systems \ 0.05, SDS: 0.63 ± 0.62 vs. 0.36 ± 0.47, P = 0.00007*, (group 1) reached a significantly higher degree of inter- P \ 0.05). Mean SRS scores were not significantly rater agreement in comparison to reader groups 2 and 3 534 Mansour et al Journal of Nuclear Cardiology Multi-center study of inter-rater reproducibility, image quality… March/April 2023 Figure 2. Image quality of conventional SPECT vs. CZT-SPECT in the groups of readers with varying experience. Bars represent mean image quality. Group 1: readers experienced with conventional SPECT and CZT-SPECT; Group 2: readers only experienced with conventional SPECT; Group 3: readers with little or no experience in interpretation of SPECT MPI. Figure 3. Mean artifacts MPI, myocardial perfusion imaging; CZT cadmium-zinc-telluride. (CZT-MPI 0.54, Anger-MPI 0.49, P \ 0.05). Group 2 group 3 was always lower compared to group 1 showed significantly lower inter-rater agreement in (Figure 5). CZT-MPI in comparison to Anger-MPI (0.23 vs. 0.37 For the assessment of interpretative certainty, the in diagnosis and 0.29 vs. 0.43 in abnormality, P\ 0.05). answers were divided into three columns: equivocal, Group 3, with little or no experience in MPI, achieved a probable (normal or abnormal), and definite (normal or significantly higher degree of agreement in CZT-MPI abnormal). Higher values indicated greater interpretative than with the Anger-MPI (0.26 vs. 0.17 in diagnosis and certainty regarding final diagnosis based purely on 0.36 vs. 0.19 in abnormality, P \ 0.05), respectively. image interpretation. Experienced readers (groups 1 and Regardless of the modality, the inter-rater agreement of 2) had significantly higher interpretative certainty than readers with little or no experience in MPI interpretation Journal of Nuclear Cardiology Mansour et al 535 Volume 30, Number 2;528–39 Multi-center study of inter-rater reproducibility, image quality… Figure 4. Scores based on 17 segment model of the American heart association (AHA). SSS summed stress score, SRS summed rest score, SDS summed difference score, MPI myocardial perfuion imaging, CZT cadmium-zinc-telluride. Figure 5. Inter-rater agreement. Inter-rater agreement was analyzed with the Fleiss’ Kappa statistics. I: Inter-rater agreement regarding all answers for diagnosis (normal, ischemia, mixed ischemia/scar and scar); II: Inter-rater agreement regarding normal versus abnormal findings (normal versus all other answers). The kappa value is shown as a point with the respective two- sided confidence interval of 95%. Interpretation: \ 0.00 = Poor agreement; 0.00 to 0.20 = Slight agreement; 0.21 to 0.40 = fair agreement; 0.41 to 0.60 = Moderate agreement; 0.61 to 0.80 = Substantial agreement; 0.81 to 1.00 = Almost perfect agreement. The red asterisk indicates statistical significance (P B 0.05). (group 3), (2.5 vs. 2.2 in CZT-MPI, 2.43 vs. 2.14 in a C 70% stenosis of at least one coronary artery, and two Anger-MPI, P \ 0.05). Mean overall diagnostic cer- patients had normal angiograms. A negative control tainty was comparable for both camera systems (2.34 vs. group (N = 10) was defined as patients with no relevant 2.39, P = 0.38). coronary artery stenosis in coronary angiography (N = 2) as well as patients who had intermediate pretest probability for CAD, yet no indication for a coronary Sensitivity and specificity angiography after revision of all non invasive examina- In patients undergoing coronary angiography, eight tions and available data and exclusion of CAD in patients had relevant coronary artery stenosis, defined as interdisciplinary consensus (N = 8). Overall, the D- 536 Mansour et al Journal of Nuclear Cardiology Multi-center study of inter-rater reproducibility, image quality… March/April 2023 Table 2. Sensitivity and specificity Camera Sensitivity (%) Specificity (%) Accuracy (%) Group 1 Conventional 37.5 100 72 CZT 62.5 60 61 Group 2 Conventional 75 80 78 CZT 87.5 30 56 Group 3 Conventional 87.5 60 72 CZT 87.5 20 50 Overall Conventional 62.5 100 83 CZT 87.5 40 61 Group 1: readers experienced with both conventional SPECT cameras and a CZT-SPECT camera; Group 2: readers only experienced with conventional SPECT cameras; Group 3: readers with little or no experience in MPS interpretation SPECT showed a higher sensitivity compared to con- reader groups, regardless of underlying experience. ventional SPECT cameras (88% versus 63%, P \0.05, When compared with coronary angiography, higher CI-CZT 0.47-0.99, CI-conv. 0.25 to 0.91). Specificity of sensitivity (87.5% vs 62.5%) and significantly lower the CZT camera was significantly lower than of the specificity for the CZT camera were observed (40% vs conventional cameras (40% versus 100%, P \ 0.05, CI- 100%). In the meta-analysis by Nudi et al.8 comparing CZT. 0.12 to 0.74, CI-conv. 0.69 to 1.00). Higher rates CZT-MPI and conventional Anger-MPI, breast attenu- of false positive evaluations resulted in overall reduced ation and obesity were discussed as possible causes for accuracy for CZT MPI, as shown in Table 2 (61% versus the reduced specificity of CZT technology as extra- 83%, P \0.05). cardiac soft tissue activity can still hamper image quality and contrast-to-noise ratio possibly generating false positive results arising especially in inferior and infer- DISCUSSION olateral segments.9,10 Some studies suggest greater Novel CZT-MPI systems have many advantages diagnostic accuracy in combined supine and prone over conventional Anger-MPI, a concept which has not CZT MPI without adversely affecting its sensitivity.25 changed significantly over more than five decades, However, this comes at the cost of extended examina- however, the influence on inter-reader reproducibility tion time and is therefore recommended in the case of and diagnostic certainty compared to conventional inferolateral ischemia in patients with a history of SPECT systems has been insufficiently investigated. previous inferolateral MI.26 Therefore, routine prone The aim of this study was to compare two technolog- imaging in CZT-MPI and adequate training and learning ically different camera systems in the identical patient curve of the technologists, especially with patient cohort with the use of a standardized questionnaire in a positioning, are key in yielding the best images this multicentric setting and to study inter-rater reproducibil- technology can provide. Lack of experience of the ity, reader experience effects, and diagnostic accuracy of technologists at the time of introduction of the D- the different MPI technologies among readers from the SPECT could have contributed to the low specificity of field of nuclear cardiology with different levels of this camera in our study. experience in MPI interpretation. Although the stress MPI with the CZT camera was Better image quality of the CZT camera was performed as first procedure in 80% of the patients, confirmed and is technically due to its significantly artifacts due to liver and bowel activity were not enhanced sensitivity of count detection compared to considered significantly more serious for the CZT conventional Anger cameras as previously confirmed in camera. It is noticeable that the experience of the reader phantom models by Zoccarato et al.24. In addition, heart- plays a significant role in the interpretation of artifacts as specific iterative reconstruction methods, considering such. All inexperienced readers (group 3) achieved a the collimation geometry of CZT-MPI, improve image significantly lower specificity with the CZT camera contrast and reduce noise level as confirmed in this (mean specificity of 20%). Overall, they observed study. The good quality of CZT images is also significantly more abnormalities in the CZT-MPI find- confirmed with less breast attenuation observed by all ings than in conventional Anger-MPI findings. The Journal of Nuclear Cardiology Mansour et al 537 Volume 30, Number 2;528–39 Multi-center study of inter-rater reproducibility, image quality… tendency for overestimation of the disease was shown being misinterpreted as perfusion abnormalities How- repeatedly in questions about diagnosis and in assess- ever, routine attenuation correction (e.g., by CT or by ments with significantly higher SSS and SDS reached by routine repositioning of the patient) is not performed in this group (Figure 4). Fiechter et al.6 discussed that due our clinic, so that attenuation-corrected images would to the improved spatial resolution of the CZT detectors, only be available for a small number of patients. better visualization of non-serious perfusion defects in Therefore, we decided to compare the images primarily the absence of significant CAD can also impair the created in the clinic with standard positioning of the specificity of CZT-based camera systems as also patient, also in order to better capture the different observed in this study. A large multicenter study image properties of the respective camera systems. evaluating fractional flow reserve (FFR) in coronary lesions found that the majority of coronary lesions with CONCLUSION a narrowing of 50–70% have no hemodynamic rele- vance.27 Although the degree of agreement between CZT-MPI delivers higher diagnostic sensitivity conventional Anger-MPI and FFR measurements has compared to conventional Angel-MPI whilst exhibiting been well investigated, to our knowledge, this is not the significantly lower specificity. Our investigation has case for CZT-MPI.28 Therefore, further investigation of shown that the reduced accuracy of CZT-MPI due to low CZT-MPI in correlation with the FFR will certainly be specificity compared to Anger-MPI—especially required in the future. achieved by nuclear cardiologist with no previous Significantly higher Inter-rater agreement in the experience in CZT-MPI—illustrates the necessity of a reader group with experience with CZT based cameras dedicated training. The importance of a differentiated (group 1) compared to the rest of the groups (2 and 3) evaluation and interpretation of the results of non- with no significant experience with CZT-MPI interpre- invasive diagnostics in everyday decision-making and tation illustrates that regardless of the underlying clinical practice was demonstrated. experience, dedicated training with such systems is required to avoid over- or underestimation of results. NEW KNOWLEDGE GAINED Importantly, for optimal image acquisition technologists should also undergo training and acquire experience CZT-MPI displays superior image quality and lower regarding optimal patient positioning in CZT-MPI.29 susceptibility for artifacts. Reader experience plays a major role in the interpretation of SPECT images, especially when acquired by camera systems with high LIMITATIONS OF THE STUDY diagnostic sensitivity. Therefore, when introducing a It is important to emphasize that a purely image- CZT-detector-based camera, dedicated training of the based, visual evaluation of the myocardial perfusion medical staff is recommended. scintigraphy can be considered as a study limitation. This does not reflect the clinical reality, as various clinical parameters play a role in the diagnostic work-up Author contributions of CAD and everyday decision-making. However, as we were interested in a direct head-to-head comparison of NM, SN, and CR contributed in the (1) conception and imaging systems, we considered this unavoidable. The design or analysis and interpretation of data or both; (2) active use of a standardized reading template, the anonymiza- involvement in collecting data or performing experiments with tion and randomization of the modality of the subsequent participation in data analysis; (3) drafting of the manuscript or revising it critically for important intellectual examination (CZT- vs. Anger-MPI) and the multicentric content; and (4) final approval of the manuscript submitted. approach helped increase the power of this study and ER, GA, PG, CA, PB, IB, AF, FF, RC, FK, LK, MP and PS reduce familiarity bias. Another limitation in the use of a contributed in the (1) conception and design or analysis and single cohort and sequential double examination is the interpretation of data or both; (2) active involvement in extra cardiac clearance pattern and the limited but collecting data or performing experiments with subsequent nevertheless existing decay of radiotracer with time participation in data analysis; and (4) final approval of the between the examinations. However, a study with manuscript submitted. separate examination protocols for the same patient cohort on different days is not justifiable due to the Funding unnecessary radiation exposure. In the present study, no attenuation corrected images (either by CT or by Open Access funding enabled and organized by Projekt alternative positioning of the patient) were provided to DEAL. This study was not supported by any funding. the readers. This came at the risk of attenuation artifacts 538 Mansour et al Journal of Nuclear Cardiology Multi-center study of inter-rater reproducibility, image quality… March/April 2023 Disclosures quantification, visual interpretation, and coronary angiography. J Nabeel Mansour, Eliana Reyes, Panagiotis Georgoulias, Nucl Cardiol 2013;20:763-73. 4. Garcia EV, Faber TL, Esteves FP. Cardiac dedicated ultrafast Constantinos Anagnostopoulos, Paco Bravo, Isabella Bruno, SPECT cameras: new designs and clinical implications. J Nucl Albert Flotats, Francisco Fuentes-Ocampo, Roberto Sciagrà, Med 2011;52:210-7. Felix Keng, Lukas Kessler, Maria Papathanasiou, Prem 5. Allie R, Hutton BF, Prvulovich E, Bomanji J, Michopoulou S, Soman, and Christoph Rischpler declares that they have no Ben-Haim S. Pitfalls and artifacts using the D-SPECT dedicated conflict of interest. cardiac camera. J Nucl Cardiol 2016;23:301-10. 6. Fiechter M, Ghadri JR, Kuest SM, Pazhenkottil AP, Wolfrum M, Ethical approval Nkoulou RN, et al. Nuclear myocardial perfusion imaging with a All procedures performed in studies involving human novel cadmium-zinc-telluride detector SPECT/CT device: first validation versus invasive coronary angiography. Eur J Nucl Med participants were in accordance with the ethical standards of Mol Imaging 2011;38:2025-30. the institutional research committee and with the 1964 7. Khoshakhlagh M, Islamian JP, Abedi SM, Mahmoudian B. Helsinki declaration and its later amendments or comparable Development of scintillators in nuclear medicine. World J Nucl ethical standards. Med. 2015;14:156-9. 8. Nudi F, Iskandrian AE, Schillaci O, Peruzzi M, Frati G, Biondi- Informed consent Zoccai G. Diagnostic accuracy of myocardial perfusion imaging For this type of study formal consent is not required. with CZT technology: Systemic review and meta-analysis of comparison with invasive coronary angiography. JACC Cardio- Consent for publication vasc Imaging 2017;10:787-94. 9. Liu C-J, Cheng J-S, Chen Y-C, Huang Y-H, Yen R-F. A perfor- Consent for publication was obtained for every individual mance comparison of novel cadmium–zinc–telluride camera and person’s data included in the study. conventional SPECT/CT using anthropomorphic torso phantom and water bags to simulate soft tissue and breast attenuation. Ann Nucl Med 2015;29:342-50. 10. Oddstig J, Martinsson E, Jögi J, Engblom H, Hindorf C. Differ- Open Access ences in attenuation pattern in myocardial SPECT between CZT and conventional gamma cameras. J Nucl Cardiol 2019;26:1984- This article is licensed under a Creative Commons 91. Attribution 4.0 International License, which permits use, 11. Esteves FP, Raggi P, Folks RD, Keidar Z, Askew JW, Rispler S, sharing, adaptation, distribution and reproduction in any et al. Novel solid-state-detector dedicated cardiac camera for fast medium or format, as long as you give appropriate credit to myocardial perfusion imaging: multicenter comparison with the original author(s) and the source, provide a link to the standard dual detector cameras. J Nucl Cardiol 2009;16:927-34. Creative Commons licence, and indicate if changes were 12. Sharir T, Ben-Haim S, Merzon K, Prochorov V, Dickman D, Ben- made. The images or other third party material in this article Haim S, et al. High-speed myocardial perfusion imaging initial are included in the article’s Creative Commons licence, unless clinical comparison with conventional dual detector anger camera indicated otherwise in a credit line to the material. If material imaging. JACC Cardiovasc Imaging 2008;1:156-63. 13. Sharir T, Slomka PJ, Hayes SW, DiCarli MF, Ziffer JA, Martin is not included in the article’s Creative Commons licence and WH, et al. Multicenter trial of high-speed versus conventional your intended use is not permitted by statutory regulation or single-photon emission computed tomography imaging: quantita- exceeds the permitted use, you will need to obtain permission tive results of myocardial perfusion and left ventricular function. J directly from the copyright holder. To view a copy of this Am Coll Cardiol 2010;55:1965-74. licence, visit http://creativecommons.org/licenses/by/4.0/. 14. Bittencourt MS, Hulten E, Polonsky TS, Hoffman U, Nasir K, Abbara S, et al. European Society of Cardiology-Recommended Coronary Artery Disease Consortium Pretest Probability Scores References More Accurately Predict Obstructive Coronary Disease and Car- diovascular Events Than the Diamond and Forrester Score: The 1. Fihn SD, Gardin JM, Abrams J, Berra K, Blankenship JC, Dallas Partners Registry. Circulation 2016;134:201-11. AP. 2012 ACCF/AHA/ACP/AATS/PCNA/SCAI/STS guideline 15. Electric Paper Evaluationssysteme GmbH. EvaSys SurveyGrid for the diagnosis and management of patients with stable ischemic 2020 [Available from: https://surveygrid.evasys.de. heart disease: a report of the American College of Cardiology 16. Berman DS, Abidov A, Kang X, Hayes SW, Friedman JD, Sci- Foundation/American Heart Association task force on practice ammarella MG, et al. Prognostic validation of a 17-segment score guidelines, and the American College of Physicians, American derived from a 20-segment score for myocardial perfusion SPECT Association for Thoracic Surgery, Preventive Cardiovascular interpretation. J Nucl Cardiol 2004;11:414-23. Nurses Association, Society for Cardiovascular Angiography and 17. Erlandsson K, Kacperski K, Van Gramberg D, Hutton BF. Per- Interventions, and Society of Thoracic Surgeons. J Am Coll Car- formance evaluation of D-SPECT: a novel SPECT system for diol 2012;60:e44-164. nuclear cardiology. Phys Med Biol 2009;54:2635. 2. Slomka PJ, Patton JA, Berman DS, Germano G. Advances in 18. Verberne HJ, Acampa W, Anagnostopoulos C, Ballinger J, Bengel technical aspects of myocardial perfusion SPECT imaging. J Nucl F, De Bondt P, et al. EANM procedural guidelines for radionuclide Cardiol 2009;16:255-76. myocardial perfusion imaging with SPECT and SPECT/CT. CT; 3. Duvall WL, Slomka PJ, Gerlach JR, Sweeny JM, Baber U, Croft 2018. LB, et al. High-efficiency SPECT MPI: comparison of automated Journal of Nuclear Cardiology Mansour et al 539 Volume 30, Number 2;528–39 Multi-center study of inter-rater reproducibility, image quality… 19. Nekolla SG. MunichHeart, A Platform For Integrated Cardiac 26. Goto K, Takebayashi H, Kihara Y, Yamane H, Hagikura A, Imaging 2010 [updated 2014. Available from: http://www.munich Morimoto Y, et al. Impact of combined supine and prone heart.de/docs/index.html. myocardial perfusion imaging using an ultrafast cardiac gamma 20. Nekolla SG, Miethaner C, Nguyen N, Ziegler SI, Schwaiger M. camera for detection of inferolateral coronary artery disease. Int J Reproducibility of polar map generation and assessment of defect Cardiol 2014;174:313-7. severity and extent assessment in myocardial perfusion imaging 27. Tonino PA, Fearon WF, De Bruyne B, Oldroyd KG, Leesar MA, using positron emission tomography. Eur J Nucl Med Ver Lee PN, et al. Angiographic versus functional severity of 1998;25:1313-21. coronary artery stenoses in the FAME study fractional flow reserve 21. Gehan EA. A generalized Wilcoxon test for comparing arbitrarily versus angiography in multivessel evaluation. J Am Coll Cardiol singly-censored samples. Biometrika 1965;52:203-24. 2010;55:2816-21. 22. Dixon WJ, Mood AM. The statistical sign test. J Am Stat Assoc 28. Barone-Rochette G, Leclere M, Calizzano A, Vautrin E, Celine 1946;41:557-66. GC, Broisat A, et al. Stress thallium-201/rest technetium-99m 23. Fleiss JL. Measuring nominal scale agreement among many raters. sequential dual-isotope high-speed myocardial perfusion imaging Psychol Bull 1971;76:378. validation versus invasive coronary angiography. J Nucl Cardiol 24. Zoccarato O, Lizio D, Savi A, Indovina L, Scabbio C, Leva L, 2015;22:513-22. et al. Comparative analysis of cadmium-zincum-telluride cameras 29. Johnson RD, Bath NK, Rinker J, Fong S, James SS, Pampaloni dedicated to myocardial perfusion SPECT: A phantom study. J MH, et al. Introduction to the D-SPECT for technologists: work- Nucl Cardiol 2016;23:885-93. flow using a dedicated digital cardiac camera. J Nucl Med Technol 25. Nishiyama Y, Miyagawa M, Kawaguchi N, Nakamura M, Kido T, 2020;48:297-303. Kurata A, et al. Combined supine and prone myocardial perfusion single-photon emission computed tomography with a cadmium Publisher’s Note Springer Nature remains neutral with regard to zinc telluride camera for detection of coronary artery disease. Circ jurisdictional claims in published maps and institutional affiliations. J 2014;78:1169-75.