Annals of Nuclear Medicine (v.32, #5)

Artificial neural network retrained to detect myocardial ischemia using a Japanese multicenter database by Kenichi Nakajima; Koichi Okuda; Satoru Watanabe; Shinro Matsuo; Seigo Kinuya; Karin Toth; Lars Edenbrandt (303-310).
An artificial neural network (ANN) has been applied to detect myocardial perfusion defects and ischemia. The present study compares the diagnostic accuracy of a more recent ANN version (1.1) with the initial version 1.0.We examined 106 patients (age, 77 ± 10 years) with coronary angiographic findings, comprising multi-vessel disease (≥ 50% stenosis) (52%) or old myocardial infarction (27%), or who had undergone coronary revascularization (30%). The ANN versions 1.0 and 1.1 were trained in Sweden (n = 1051) and Japan (n = 1001), respectively, using 99mTc-methoxyisobutylisonitrile myocardial perfusion images. The ANN probabilities (from 0.0 to 1.0) of stress defects and ischemia were calculated in candidate regions of abnormalities. The diagnostic accuracy was compared using receiver-operating characteristics (ROC) analysis and the calculated area under the ROC curve (AUC) using expert interpretation as the gold standard.Although the AUC for stress defects was 0.95 and 0.93 (p = 0.27) for versions 1.1 and 1.0, respectively, that for detecting ischemia was significantly improved in version 1.1 (p = 0.0055): AUC 0.96 for version 1.1 (sensitivity 87%, specificity 96%) vs. 0.89 for version 1.0 (sensitivity 78%, specificity 97%). The improvement in the AUC shown by version 1.1 was also significant for patients with neither coronary revascularization nor old myocardial infarction (p = 0.0093): AUC = 0.98 for version 1.1 (sensitivity 88%, specificity 100%) and 0.88 for version 1.0 (sensitivity 76%, specificity 100%). Intermediate ANN probability between 0.1 and 0.7 was more often calculated by version 1.1 compared with version 1.0, which contributed to the improved diagnostic accuracy. The diagnostic accuracy of the new version was also improved in patients with either single-vessel disease or no stenosis (n = 47; AUC, 0.81 vs. 0.66 vs. p = 0.0060) when coronary stenosis was used as a gold standard.The diagnostic ability of the ANN version 1.1 was improved by retraining using the Japanese database, particularly for identifying ischemia.
Keywords: Nuclear cardiology; Artificial intelligence; Myocardial perfusion imaging; Coronary artery disease

CT-based attenuation correction and resolution compensation for I-123 IMP brain SPECT normal database: a multicenter phantom study by Yoshitaka Inui; Takashi Ichihara; Masaki Uno; Masanobu Ishiguro; Kengo Ito; Katsuhiko Kato; Hajime Sakuma; Hidehiko Okazawa; Hiroshi Toyama (311-318).
Statistical image analysis of brain SPECT images has improved diagnostic accuracy for brain disorders. However, the results of statistical analysis vary depending on the institution even when they use a common normal database (NDB), due to different intrinsic spatial resolutions or correction methods. The present study aimed to evaluate the correction of spatial resolution differences between equipment and examine the differences in skull bone attenuation to construct a common NDB for use in multicenter settings.The proposed acquisition and processing protocols were those routinely used at each participating center with additional triple energy window (TEW) scatter correction (SC) and computed tomography (CT) based attenuation correction (CTAC). A multicenter phantom study was conducted on six imaging systems in five centers, with either single photon emission computed tomography (SPECT) or SPECT/CT, and two brain phantoms. The gray/white matter I-123 activity ratio in the brain phantoms was 4, and they were enclosed in either an artificial adult male skull, 1300 Hounsfield units (HU), a female skull, 850 HU, or an acrylic cover. The cut-off frequency of the Butterworth filters was adjusted so that the spatial resolution was unified to a 17.9 mm full width at half maximum (FWHM), that of the lowest resolution system. The gray-to-white matter count ratios were measured from SPECT images and compared with the actual activity ratio. In addition, mean, standard deviation and coefficient of variation images were calculated after normalization and anatomical standardization to evaluate the variability of the NDB.The gray-to-white matter count ratio error without SC and attenuation correction (AC) was significantly larger for higher bone densities (p < 0.05). The count ratio error with TEW and CTAC was approximately 5% regardless of bone density. After adjustment of the spatial resolution in the SPECT images, the variability of the NDB decreased and was comparable to that of the NDB without correction.The proposed protocol showed potential for constructing an appropriate common NDB from SPECT images with SC, AC and spatial resolution compensation.
Keywords: Brain perfusion SPECT; Scatter correction; Attenuation correction; Statistical image analysis; Normal database

Diagnostic value of quantitative assessment of cardiac 18F-fluoro-2-deoxyglucose uptake in suspected cardiac sarcoidosis by Adrien Lebasnier; Damien Legallois; Boris Bienvenu; Emmanuel Bergot; Cédric Desmonts; Gérard Zalcman; Denis Agostini; Alain Manrique (319-327).
The identification of cardiac sarcoidosis is challenging as there is no gold standard consensually admitted for its diagnosis. The aim of this study was to evaluate the diagnostic value of the assessment of cardiac dynamic 18F-fluoro-2-deoxyglucose positron emission tomography (18F-FDG PET/CT) and net influx constant (Ki) in patients suspected of cardiac sarcoidosis.Data obtained from 30 biopsy-proven sarcoidosis patients suspected of cardiac sarcoidosis who underwent a 50-min list-mode cardiac dynamic 18F-FDG PET/CT after a 24 h high-fat and low-carbohydrate diet were analyzed. A normalized coefficient of variation of quantitative glucose influx constant, calculated as the ratio: standard deviation of the segmental Ki (min−1)/global Ki (min−1) was determined using a validated software (Carimas® 2.4, Turku PET Centre). Cardiac sarcoidosis was diagnosed according to the Japanese Ministry of Health and Welfare criteria. Receiving operating curve analysis was performed to determine sensitivity and specificity of cardiac dynamic 18F-FDG PET/CT analysis to diagnose cardiac sarcoidosis.Six out of 30 patients (20%) were diagnosed as having cardiac sarcoidosis. Myocardial glucose metabolism was significantly heterogeneous in patients with cardiac sarcoidosis who showed significantly higher normalized coefficient of variation values compared to patients without cardiac sarcoidosis (0.513 ± 0.175 vs. 0.205 ± 0.081; p = 0.0007). Using ROC curve analysis, we found a cut-off value of 0.38 for the diagnosis of cardiac sarcoidosis with a sensitivity of 100% and a specificity of 91%.Our results suggest that quantitative analysis of cardiac dynamic 18F-FDG PET/CT could be a useful tool for the diagnosis of cardiac sarcoidosis.
Keywords: Cardiac sarcoidosis; Positron emission tomography; Dynamic acquisition; FDG; Quantitative analysis

We evaluated the significance of breast-specific gamma imaging (BSGI) for determination of upstaging to invasive cancer from preoperative ductal carcinoma in situ (DCIS).This study enrolled 168 patients with 175 breast lesions diagnosed as DCIS on core-needle biopsy that subsequently underwent preoperative BSGI between September 2011 and October 2017. Both qualitative and quantitative analyses using tumor-to-normal background ratio (TNR) and coefficient of variation (COV) were performed, and the predictive significance for upstaging to invasive cancer was investigated. We also sought to identify clinicopathological factors associated with upstaging and their relation to BSGI findings.Fifty-eight lesions (33.1%) were confirmed to be invasive breast cancer after the final surgical approach. On univariate analysis, tumor size based on ultrasonography (US), hormone status, histologic grade, Breast Imaging Reporting and Data System category, comedo-necrosis, Ki-67 expression, and BSGI findings were associated with upstaging to invasive cancer. Tumor size > 2.0 cm (p = 0.005), Ki-67 expression > 8% (p < 0.001), qualitative BSGI findings (p = 0.020), and COV > 30.44 (p = 0.022) were independently associated with upstaging after multiple regression analysis.BSGI is a useful imaging modality for predicting upstaging to invasive breast cancer from DCIS on core-needle biopsy in conjunction with US tumor size and Ki-67 expression.
Keywords: Breast-specific gamma imaging; Ductal carcinoma in situ; Coefficient of variation; Invasiveness; Heterogeneity

Statistical SPECT reconstruction can be very time-consuming especially when compensations for collimator and detector response, attenuation, and scatter are included in the reconstruction. This work proposes an accelerated SPECT reconstruction algorithm based on graphics processing unit (GPU) processing.Ordered subset expectation maximization (OSEM) algorithm with CT-based attenuation modelling, depth-dependent Gaussian convolution-based collimator-detector response modelling, and Monte Carlo-based scatter compensation was implemented using OpenCL. The OpenCL implementation was compared against the existing multi-threaded OSEM implementation running on a central processing unit (CPU) in terms of scatter-to-primary ratios, standardized uptake values (SUVs), and processing speed using mathematical phantoms and clinical multi-bed bone SPECT/CT studies.The difference in scatter-to-primary ratios, visual appearance, and SUVs between GPU and CPU implementations was minor. On the other hand, at its best, the GPU implementation was noticed to be 24 times faster than the multi-threaded CPU version on a normal 128 × 128 matrix size 3 bed bone SPECT/CT data set when compensations for collimator and detector response, attenuation, and scatter were included.GPU SPECT reconstructions show great promise as an every day clinical reconstruction tool.
Keywords: SPECT reconstruction; Scatter correction; Monte Carlo; Graphics processing unit (GPU)

Diagnostic performance of 18F-FDG PET/CT and whole-body diffusion-weighted imaging with background body suppression (DWIBS) in detection of lymph node and bone metastases from pediatric neuroblastoma by Hiroaki Ishiguchi; Shinji Ito; Katsuhiko Kato; Yusuke Sakurai; Hisashi Kawai; Naotoshi Fujita; Shinji Abe; Atsushi Narita; Nobuhiro Nishio; Hideki Muramatsu; Yoshiyuki Takahashi; Shinji Naganawa (348-362).
Recent many studies have shown that whole body “diffusion-weighted imaging with background body signal suppression” (DWIBS) seems a beneficial tool having higher tumor detection sensitivity without ionizing radiation exposure for pediatric tumors. In this study, we evaluated the diagnostic performance of whole body DWIBS and 18F-FDG PET/CT for detecting lymph node and bone metastases in pediatric patients with neuroblastoma.Subjects in this retrospective study comprised 13 consecutive pediatric patients with neuroblastoma (7 males, 6 females; mean age, 2.9 ± 2.0 years old) who underwent both 18F-FDG PET/CT and whole-body DWIBS. All patients were diagnosed as neuroblastoma on the basis of pathological findings. Eight regions of lymph nodes and 17 segments of skeletons in all patients were evaluated. The images of 123I-MIBG scintigraphy/SPECT-CT, bone scintigraphy/SPECT, and CT were used to confirm the presence of lymph node and bone metastases. Two radiologists trained in nuclear medicine evaluated independently the uptake of lesions in 18F-FDG PET/CT and the signal-intensity of lesions in whole-body DWIBS visually. Interobserver difference was overcome through discussion to reach a consensus. The sensitivities, specificities, and overall accuracies of 18F-FDG PET/CT and whole-body DWIBS were compared using McNemer’s test. Positive predictive values (PPVs) and negative predictive values (NPVs) of both modalities were compared using Fisher’s exact test.The total numbers of lymph node regions and bone segments which were confirmed to have metastasis in the total 13 patients were 19 and 75, respectively. The sensitivity, specificity, overall accuracy, PPV, and NPV of 18F-FDG PET/CT for detecting lymph node metastasis from pediatric neuroblastoma were 100, 98.7, 98.9, 95.0, and 100%, respectively, and those for detecting bone metastasis were 90.7, 73.1, 80.3, 70.1, and 91.9%, respectively. In contrast, the sensitivity, specificity, overall accuracy, PPV, and NPV of whole-body DWIBS for detecting bone metastasis from pediatric neuroblastoma were 94.7, 24.0, 53.0, 46.4 and 86.7%, respectively, whereas those for detecting lymph node metastasis were 94.7, 85.3, 87.2, 62.1, and 98.5%, respectively. The low specificity, overall accuracy, and PPV of whole-body DWIBS for detecting bone metastasis were due to a high incidence of false-positive findings (82/108, 75.9%). The specificity, overall accuracy, and PPV of whole-body DWIBS for detecting lymph node metastasis were also significantly lower than those of 18F-FDG PET/CT for detecting lymph node metastasis, although the difference between these 2 modalities was less than that for detecting bone metastasis.The specificity, overall accuracy, and PPV of whole-body DWIBS are significantly lower than those of 18F-FDG PET/CT because of a high incidence of false-positive findings particularly for detecting bone metastasis, whereas whole-body DWIBS shows a similar level of sensitivities for detecting lymph node and bone metastases to those of 18F-FDG PET/CT. DWIBS should be carefully used for cancer staging in children because of its high incidence of false-positive findings in skeletons.
Keywords: 18F-FDG PET/CT; Whole-body DWIBS; Neuroblastoma; Metastasis

Quantitative evaluation of the tracer distribution in dopamine transporter SPECT for objective interpretation by Yu Iwabuchi; Tadaki Nakahara; Masashi Kameyama; Yoshitake Yamada; Masahiro Hashimoto; Yuji Ogata; Yohji Matsusaka; Mari Katagiri; Kazunari Itoh; Takashi Osada; Daisuke Ito; Hajime Tabuchi; Masahiro Jinzaki (363-371).
Quantification of the tracer distribution would add objectivity to the visual assessments of dopamine transporter (DAT) single photon emission computed tomography (SPECT) data. Our study aimed to evaluate the diagnostic utility of fractal dimension (FD) as a quantitative indicator of tracer distribution and compared with the conventional quantitative value: specific binding ratio (SBR). We also evaluated the utility of the combined index SBR/FD (SBR divided by FD).We conducted both clinical and phantom studies. In the clinical study, 150 patients including 110 patients with Parkinsonian syndrome (PS) and 40 without PS were enrolled. In the phantom study, we used a striatal phantom with the striatum chamber divided into two spaces, representing the caudate nucleus and putamen. The SBR, FD, and SBR/FD were calculated and compared between datasets for evaluating the diagnostic utility. Mann–Whitney test and receiver-operating characteristics (ROC) analysis were used for analysis.ROC analysis revealed that the FD value had high diagnostic performance [the areas under the curve (AUC) = 0.943] and the combined use of SBR and FD (SBR/FD) delivered better results than the SBR alone (AUC, 0.964 vs 0.899; p < 0.001). The sensitivity, specificity, and accuracy, respectively, were 79.1, 85.0, and 80.7% with SBR, 84.5, 97.5, and 88.0% with FD, and 92.7, 87.5, and 91.3% with SBR/FD.Our results confirmed that the FD value is a useful diagnostic index, which reflects the tracer distribution in DAT SPECT images. The combined use of SBR and FD was more useful than either used alone.
Keywords: 123I-Ioflupane; 123I-FP-CIT; DAT; Fractal analysis; Fractal dimension

FDG-PET/CT at the end of immuno-chemotherapy in follicular lymphoma: the prognostic role of the ratio between target lesion and liver SUVmax (rPET) by Salvatore Annunziata; Annarosa Cuccaro; Maria Chiara Tisi; Stefan Hohaus; Vittoria Rufini (372-377).
To retrospectively investigate the prognostic role of the ratio between target lesion and liver SUVmax (rPET) in patients with follicular lymphoma (FL) submitted to FDG-PET/CT at the end of immuno-chemotherapy (PI-PET), and to compare rPET with International Harmonization Project criteria (IHP), Deauville Score (5p-DS) and FL International Prognostic Index at diagnosis (FLIPI).Eighty-nine patients with FL undergoing PI-PET were evaluated. The receiver operating characteristic (ROC) approach was applied to identify the optimal cut-point of rPET with respect to 5-years progression free survival (PFS). The prognostic significance of rPET was compared with IHP, DS and FLIPI. Positive predictive value (PPV) and negative predictive value (NPV) were calculated using the presence of adverse events as gold standard.The ROC analysis for rPET as predictor of progression showed an optimal rPET cut-point of 0.98. Patients with positive values of IHP, DS and rPET had a PFS of 50, 30 and 31%. PPV were of 56, 80 and 80%, NPV of 83, 86 and 88%, respectively. DS and rPET differed only in two patients. FLIPI was not predictive of progression and relapse.rPET is a prognostic factor in patients with FL submitted to PI-PET. Although it has a similar prognostic power as DS, it can have methodological advantages over visual analysis. PI-PET with different evaluation systems has a stronger prognostic power than FLIPI at diagnosis, so it could be useful to identify patients with FL at risk for early relapse after immuno-chemotherapy.
Keywords: Follicular lymphoma; FDG PET/CT; rPET; Deauville Score; IHP; FLIPI