Diagnostic performance of a deep-learning model using 18F-FDG PET/CT for evaluating recurrence after radiation therapy in patients with lung cancer.

OBJECTIVE: We developed a deep learning model for distinguishing radiation therapy (RT)-related changes and tumour recurrence in patients with lung cancer who underwent RT, and evaluated its performance. METHODS: We retrospectively recruited 308 patients with lung cancer with RT-related changes observed on 18F-fluorodeoxyglucose positron emission tomography-computed tomography (18F-FDG PET/CT) performed after RT. Patients were labelled as positive or negative for tumour recurrence through histologic diagnosis or clinical follow-up after 18F-FDG PET/CT. A two-dimensional (2D) slice-based convolutional neural network (CNN) model was created with a total of 3329 slices as input, and performance was evaluated with five independent test sets. RESULTS: For the five independent test sets, the area under the curve (AUC) of the receiver operating characteristic curve, sensitivity, and specificity were in the range of 0.98-0.99, 95-98%, and 87-95%, respectively. The region determined by the model was confirmed as an actual recurred tumour through the explainable artificial intelligence (AI) using gradient-weighted class activation mapping (Grad-CAM). CONCLUSION: The 2D slice-based CNN model using 18F-FDG PET imaging was able to distinguish well between RT-related changes and tumour recurrence in patients with lung cancer.

One-Year Longitudinal Changes in Tau Accumulation on [18F]PI-2620 PET in the Alzheimer Spectrum.

We investigated the longitudinal changes in cortical tau accumulation and their association with cognitive decline in patients in the Alzheimer disease (AD) continuum using 2-(2-([18F]fluoro)pyridin-4-yl)-9H-pyrrolo[2,3-b:4,5c']dipyridine ([18F]PI-2620) PET. Methods: We prospectively enrolled 52 participants (age, 69.7 ± 8.4 y; 18 men and 34 women): 7 with normal cognition, 28 with mild cognitive impairment, and 17 with AD. They all completed the [18F]PI-2620 and [18F]florbetaben PET, MRI, and neuropsychologic tests at baseline and, excepting the [18F]florbetaben PET, at the 1-y follow-up. Amyloid-ß (Aß) PET images were visually scored as positive (+) or negative (-). Patients on the AD continuum, including Aß+ mild cognitive impairment and AD, were classified into early-onset (EO+) (<65 y old) or late-onset (LO+) (≥65 y old) groups. [18F]PI-2620 PET SUV ratios (SUVRs) were determined by calculating the cerebral-to-inferior cerebellar ratio. Cortical volumes were calculated using 3-dimensional T1-weighted MRI. The correlation between tau accumulation progression and cognitive decline was also investigated. Results: The global [18F]PI-2620 PET SUVRs were 1.04 ± 0.07 in 15 Aß- patients, 1.18 ± 0.21 in 20 LO+ patients (age, 76.7 ± 3.8 y), and 1.54 ± 0.38 in 17 EO+ patients (age, 63.4 ± 5.4 y; P < 0.001) at baseline. The global SUVR increased over 1 y by 0.05 ± 0.07 (3.90%) and 0.13 ± 0.22 (8.41%) in the LO+ and EO+ groups, respectively, whereas in the Aß- groups, it remained unchanged. The EO+ group showed higher global and regional tau deposition than did the Aß- and LO+ groups (P < 0.05 for each) and rapid accumulation in Braak stage V (0.15 ± 0.25; 9.10% ± 12.27%; P = 0.016 and 0.008), Braak stage VI (0.08 ± 0.12; 7.16% ± 10.06%; P < 0.006 and 0.005), and global SUVR (P = 0.013) compared with the Aß- group. In the EO+ group, the changes in SUVR in Braak stages II-VI were strongly correlated with the baseline and changes in verbal memory (P < 0.03). The LO+ group showed higher tau accumulation in Braak stage I-IV areas than did the Aß- group (P < 0.001 for each). In the LO+ group, the change in SUVR in Braak stages III and IV moderately correlated with the change in attention (P < 0.05), and the change in SUVR in Braak stages V and VI moderately correlated with the change in visuospatial function (P < 0.005). Conclusion: These findings suggest that [18F]PI-2620 PET can be a biomarker to provide regional and chronologic information about tau pathology in the AD continuum.

Prediction of [177Lu]Lu-DOTA-TATE therapy response using the absorbed dose estimated from [177Lu]Lu-DOTA-TATE SPECT/CT in patients with metastatic neuroendocrine tumour.

BACKGROUND: Peptide receptor radionuclide therapy (PRRT) with [177Lu]Lu-DOTA-TATE has shown efficacy in patients with metastatic neuroendocrine tumours (NETs). Personalised dosimetry is crucial to optimise treatment outcomes and minimise adverse events. In this study, we investigated the correlation between the tumour-absorbed dose (TAD) estimated from [177Lu]Lu-DOTA-TATE SPECT/CT and the therapeutic response. METHOD: A retrospective analysis was conducted on patients with advanced well-differentiated NETs grades 1-3 who underwent PRRT and exhibited greater uptake than liver on pre-therapeutic [68Ga]Ga-DOTA-TOC PET/CT. Target lesions were selected based on the RECIST 1.1 and PERCIST 1.0 criteria using [177Lu]Lu-DOTA-TATE SPECT/CT and pre-therapeutic contrast-enhanced CT scans. For anatomical image analysis, the sum of the longest diameter (SLD) of the target lesions was measured using the RECIST 1.1 criteria for patient-based analysis and the longest diameter (LD) of the target lesion using the RECIST-L criteria for lesion-based analysis. Standardised uptake values (SUVs) were measured on SPECT/CT images, and TADs were calculated based on the SUVs. Dosimetry was performed using a single SPECT/CT imaging time point at day 4-5 post-therapy. Statistical analyses were conducted to investigate correlations and determine the target lesion responses. RESULTS: Twenty patients with primary tumour sites and hepatic metastases were included. Fifty-five target lesions, predominantly located in the pancreas and liver, were analysed. The cumulative TAD (lesion-based analysis: r = 0.299-0.301, p = 0.025-0.027), but not the cycle 1 SUV (lesion-based analysis: r = 0.198-0.206, p = 0.131-0.147) or cycle 1 TAD (lesion-based analysis: r = 0.209-0.217, p = 0.112-0.126), exhibited a significant correlation with the change in LD of the target lesion. Binary logistic regression analysis identified the significance of the cumulative TAD in predicting disease control according to the RECIST-L criteria (odds ratio = 1.031-1.051, p = 0.024-0.026). CONCLUSIONS: The cumulative TAD estimated from [177Lu]Lu-DOTA-TATE SPECT/CT revealed a significant correlation with change in LD, which was significantly higher for the cumulative TAD than for the cycle 1 SUV or TAD. A higher cumulative TAD was associated with disease control in the target lesion. However, considering the limitations inherent to a confined sample size, careful interpretation of these findings is required. Estimation of the cumulative TAD of [177Lu]Lu-DOTA-TATE therapy could guide the platform towards personalised therapy.

[18F]fluorodeoxyglucose positron emission tomography/computed tomography characteristics of primary mediastinal germ cell tumors.

Primary mediastinal germ cell tumor (MGCT) is an uncommon tumor. Although it has histology similar to that of gonadal germ cell tumor (GCT), the prognosis for MGCT is generally worse than that for gonadal GCT. We performed visual assessment and quantitative analysis of [18F]fluorodeoxyglucose positron emission tomography/computed tomography ([18F]FDG PET/CT) for MGCTs. A total of 35 MGCT patients (age = 33.1 ± 16.8 years, F:M = 16:19) who underwent preoperative PET/CT were retrospectively reviewed. The pathologic diagnosis of MGCTs identified 24 mature teratomas, 4 seminomas, 5 yolk sac tumors, and 2 mixed germ cell tumors. Visual assessment was performed by categorizing the uptake intensity, distribution, and contour of primary MGCTs. Quantitative parameters including the maximum standardized uptake value (SUVmax), tumor-to-background ratio (TBR), metabolic tumor volume (MTV), total lesion glycolysis (TLG), and maximum diameter were compared between benign and malignant MGCTs. On visual assessment, the uptake intensity was the only significant parameter for differentiating between benign and malignant MGCTs (p = 0.040). In quantitative analysis, the SUVmax (p < 0.001), TBR (p < 0.001), MTV (p = 0.033), and TLG (p < 0.001) showed significantly higher values for malignant MGCTs compared with benign MGCTs. In receiver operating characteristic (ROC) curve analysis of these quantitative parameters, the SUVmax had the highest area under the curve (AUC) (AUC = 0.947, p < 0.001). Furthermore, the SUVmax could differentiate between seminomas and nonseminomatous germ cell tumors (p = 0.042) and reflect serum alpha fetoprotein (AFP) levels (p = 0.012). The visual uptake intensity and SUVmax on [18F]FDG PET/CT showed discriminative ability for benign and malignant MGCTs. Moreover, the SUVmax may associate with AFP levels.

Performance of deep learning models for response evaluation on whole-body bone scans in prostate cancer.

OBJECTIVE: We aimed to develop deep learning classifiers for assessing therapeutic response on bone scans of patients with prostate cancer. METHODS: A set of 3791 consecutive bone scans coupled with their last previous scan (1528 patients) was evaluated. Bone scans were labeled as "progression" or "nonprogression" on the basis of clinical reports and image review. A 2D-convolutional neural network architecture was trained with three different preprocessing methods: 1) no preprocessing (Raw), 2) spatial normalization (SN), and 3) spatial and count normalization (SCN). Data were allocated into training, validation, and test sets in the ratio of 72:8:20, with the 20% independent test set rotating all scans over a five-fold testing procedure. A Grad-CAM algorithm was employed to generate class activation maps to visualize the lesions contributing to the decision. Diagnostic performance was compared using area under the receiver operating characteristics curves (AUCs). RESULTS: The data consisted of 791 scans labeled as "progression" and 3000 scans labeled as "nonprogression." The AUCs of the classifiers were 0.632-0.710 on the Raw dataset, were significantly higher with the use of SN at 0.784-0.854 (p < 0.001 for Raw versus SN), and higher still with SCN at 0.954-0.979 (p < 0.001 for SN versus SCN). Class activation maps of the SCN model visualized lesions contributing to the model's decision of progression. CONCLUSION: With preprocessing of spatial and count normalization, our deep learning model achieved excellent performance in classifying the therapeutic response of bone scans in patients with prostate cancer.

Scaling problem in Parkinson's disease patients with pain.

INTRODUCTION: Although pain is common in Parkinson's disease (PD), the underlying mechanism remains unknown. Scaling function and dopaminergic hypofunction may contribute to pain development because increased pain sensitivity is observed in PD and is normalized after levodopa administration. We aimed to determine whether spatial discrimination (SD) and striatal dopaminergic activity (DA) differed between PD patients with and without pain. METHODS: We divided 90 patients with drug-naïve PD into two groups based on the presence or absence of pain and compared the SD threshold (SDT). We evaluated the correlation of the SDT with pain severity in PD with pain. We also compared the DA of 48 patients and analyzed the correlation with pain severity in PD patients with pain. RESULTS: The SDTs did not differ between the two groups, but unmeasurable SDT was more frequent in PD with pain. There was a positive correlation of pain severity with the SDT of the more affected hand but no correlation with the SDT of the less affected hand. The DA did not differ between the groups. There was a negative trend of pain severity with the DA of the ventral striatum (VS) but no correlation with the other striatal subregions. CONCLUSIONS: Pain in PD may be associated with scaling dysfunction in the sensory system. The abnormal scaling function would render the PD patient hypersensitive to even mild pain. The dopamine in the VS appears to be associated with pain severity; however, the relationship of striatal dopaminergic deficits with pain occurrence requires further investigation.

Diagnostic accuracy of 18F-FP-CIT PET for clinically uncertain Parkinsonian syndrome.

18F-FP-CIT is a high-resolution imaging marker of nigrostriatal neuronal integrity, differentiating Parkinsonism with loss of dopaminergic terminals (presynaptic Parkinsonian syndrome [PS]) from Parkinsonism without nigrostriatal degeneration (non-PS). We assessed the diagnostic accuracy of 18F-FP-CIT PET in patients with clinically uncertain PS (CUPS) at the first visit. Among the 272 patients who underwent 18F-FP-CIT PET imaging at the first visit between September 2008 and July 2012, 111 had CUPS (age, 62.6 ± 10.5 y; male:female, 45:66; symptom duration, 13.1 ± 8.8 months). Uncertainty criteria included only one of the three cardinal signs of Parkinsonism, two signs without bradykinesia, or atypical signs. The baseline clinical and 18F-FP-CIT PET imaging diagnostic accuracy was compared with the accuracy of clinical diagnosis after > 2-year follow-up. Nuclear medicine physicians assessed the 18F-FP-CIT PET images visually. Focal dopamine transporter binding deficit in the posterior putamen was considered PS. Bilateral symmetric striatum without focal deficit, suggesting normal 18F-FP-CIT PET, and focal deficits elsewhere in the striatum suggesting vascular Parkinsonism were considered non-PS. Seventy-nine patients had PS, and 32 did not. Baseline clinical diagnosis included PS in 45 patients, non-PS in 24, and inconclusive in 42. Among patients in whom initial clinical diagnosis (PS or non-PS) was possible, the sensitivity, specificity, and accuracy of the baseline clinical and 18F-FP-CIT PET imaging diagnoses were 54.4, 50.0, and 53.2%, and 98.7, 100, and 99.1%, respectively. The respective positive and negative predictive values were 95.6 and 66.7%, and 100 and 97.0%. Among those with initially inconclusive diagnosis, 64.2% were eventually diagnosed with PS while 35.7% were diagnosed with non-PS. The final clinical diagnosis of these patients all matched those made by 18F-FP-CIT PET imaging, except in one patient with scan without evidence of dopaminergic deficit (SWEDD). 18F-FP-CIT PET diagnosis was more accurate than clinical diagnosis, reducing the false-negative and inconclusive clinical diagnosis rates at baseline in patients with CUPS.

Retinal Thickness and Its Interocular Asymmetry Between Parkinson's Disease and Drug-Induced Parkinsonism.

BACKGROUND: Drug-induced parkinsonism (DIP) is common, but diagnosis is challenging. Although dopamine transporter imaging is useful, the cost and inconvenience are problematic, and an easily accessible screening technique is needed. We aimed to determine whether optical coherence tomography (OCT) findings could differentiate DIP from Parkinson's disease (PD). METHODS: We investigated 97 de novo PD patients and 27 DIP patients using OCT and [18F] N-(3-fluoropropyl)-2b-carbon ethoxy-3b-(4-iodophenyl) nortropane (FP-CIT) positron emission tomography. We compared peripapillary retinal nerve fiber layer thickness (pRNFLT) and macular retinal thickness (mRT) between PD and DIP patients as well as interocular differences in the pRNFLT and the mRT. Asymmetric index (%) for retinal thickness (AIRT) was calculated to measure the interocular differences between pRNFLT and mRT. The correlation between AIRT and total striatal specific/non-specific binding ratio asymmetry index (SNBRAI) was investigated in PD and DIP patients. RESULTS: No significant differences in pRNFLT and mRT values were observed between PD and DIP patients (all P values > 0.090). The mean SNBRAI was significantly higher in PD than in DIP (P = 0.008) patients; however, AIRT did not differ between PD and DIP patients in pRNFLT and mRT (all P values > 0.100). SNBRAI did not correlate with AIRT of pRNFL or mRT in PD and DIP patients (all P values > 0.060). CONCLUSION: Our study showed no benefit of retinal thickness and interocular asymmetry measurements using OCT for distinguishing PD from DIP in the early stages. Additional investigations are needed for confirmation.

MR Template-Based Individual Brain PET Volumes-of-Interest Generation Neither Using MR nor Using Spatial Normalization.

For more anatomically precise quantitation of mouse brain PET, spatial normalization (SN) of PET onto MR template and subsequent template volumes-of-interest (VOIs)-based analysis are commonly used. Although this leads to dependency on the corresponding MR and the process of SN, routine preclinical/clinical PET images cannot always afford corresponding MR and relevant VOIs. To resolve this issue, we propose a deep learning (DL)-based individual-brain-specific VOIs (i.e., cortex, hippocampus, striatum, thalamus, and cerebellum) directly generated from PET images using the inverse-spatial-normalization (iSN)-based VOI labels and deep convolutional neural network model (deep CNN). Our technique was applied to mutated amyloid precursor protein and presenilin-1 mouse model of Alzheimer's disease. Eighteen mice underwent T2-weighted MRI and 18F FDG PET scans before and after the administration of human immunoglobin or antibody-based treatments. To train the CNN, PET images were used as inputs and MR iSN-based target VOIs as labels. Our devised methods achieved decent performance in terms of not only VOI agreements (i.e., Dice similarity coefficient) but also the correlation of mean counts and SUVR, and CNN-based VOIs was highly accordant with ground-truth (the corresponding MR and MR template-based VOIs). Moreover, the performance metrics were comparable to that of VOI generated by MR-based deep CNN. In conclusion, we established a novel quantitative analysis method both MR-less and SN-less fashion to generate individual brain space VOIs using MR template-based VOIs for PET image quantification. Supplementary Information: The online version contains supplementary material available at 10.1007/s13139-022-00772-4.

Prognostic significance of pretreatment 18F-fluorodeoxyglucose positron emission tomography/computed tomography in patients with T2N1 hormone receptor-positive, ERBB2-negative breast cancer who underwent adjuvant chemotherapy.

PURPOSE: To determine whether tumor uptake of 18F-fluorodeoxyglucose (18F-FDG) is associated with invasive disease-free survival (IDFS) in patients with hormone receptor (HR)-positive ERBB2-negative early-stage breast cancer treated with adjuvant chemotherapy. METHODS: This is a single-center cohort study of women with breast cancer who underwent surgery between 2008 and 2015 at Asan Medical Center, Seoul, Korea. Patients were enrolled if they were diagnosed with HR-positive ERBB2-negative breast cancer with histology of invasive ductal carcinoma, had an American Joint Committee on Cancer pathologic tumor stage of T2N1 with 1-3 positive axillary nodes, underwent preoperative 18F-FDG positron emission tomography/computed tomography (PET/CT), and underwent breast cancer surgery followed by anthracycline- or taxane-based adjuvant chemotherapy. The primary outcome measure was IDFS. The maximum standardized uptake value (SUVmax) was dichotomized using a predefined cut-off of 4.14. RESULTS: A total of 129 patients were included. The median follow-up period for IDFS in those without recurrence was 82 months (interquartile range, 65-106). Multivariable Cox analysis showed that SUVmax was independently associated with IDFS [adjusted hazard ratio 2.49; 95% confidence interval (CI), 1.06-5.84]. Ten-year IDFS estimates via the Kaplan-Meier method were 0.60 (95% CI, 0.42-0.74) and 0.82 (95% CI, 0.65-0.91) for high and low SUVmax groups, respectively. The overall association between SUVmax and IDFS appeared to be consistent across subgroups divided according to age, progesterone receptor status, histologic grade, or presence of lymphovascular invasion. CONCLUSION: High SUVmax on preoperative 18F-FDG PET/CT was independently associated with reduced long-term IDFS in T2N1 HR-positive ERBB2-negative breast cancer patients who underwent adjuvant chemotherapy.

Correlations with REM sleep behavior disorder severity in isolated rapid eye movement sleep behavior disorders patients.

OBJECTIVES: The pathogenesis of isolated rapid eye movement sleep behavior disorders (iRBD) is poorly understood. The severity of RBD may reflect its pathogenesis. METHODS: We compared motor function and non-motor symptoms (NMSs) between iRBD patients and healthy volunteers. We correlated motor function, NMSs, and striatal dopaminergic activity with RBD severity using video-polysomnography. RESULTS: Twenty-one iRBD patients and 17 controls participated. The Unified Parkinson's Disease Rating Scale part III scores were higher in patients compared to controls (p < 0.001). There was no difference in upper extremity function between patients and controls (right, p = 0.220; left, p = 0.209), but gait was slower in iRBD patients (walking time, p < 0.001; number of steps, p < 0.001). The mean value of the Korean version of the Mini-Mental State Exam and Clinical Dementia Rating were lower in patients (p = 0.006, p = 0.003, respectively). Patients with were also more depressed (p = 0.002), had decreased olfactory function (p < 0.001), reported more frequent sleep/fatigue episodes (p < 0.001), worse attention/memory capacity (p < 0.001), gastrointestinal problems (p = 0.009), urinary problems (p = 0.007), and pain (p = 0.083). Further, iRBD patients reported more frequent sleep-related disturbances (p = 0.004), but no difference in daytime sleepiness (p = 0.663). Disease severity was correlated with pain (r = 0.686, p = 0.002) and visuospatial function (r= -0.507, p = 0.038). There were no correlations between RBD severity and striatal dopaminergic activities (p > 0.09). CONCLUSIONS: iRBD is a multisystem neurodegenerative disorder, and gait abnormalities may be a disease characteristic, possibly related to the akinetic-rigid phenotype of Parkinson's disease. The correlation between pain/visuospatial dysfunction and RBD severity may be related to its pathogenesis.

Preliminary Study for Quantitative Assessment of Sacroiliitis Activity Using Bone SPECT/CT: Comparison of Diagnostic Performance of Quantitative Parameters.

Purpose: We compared the feasibility of quantitative analysis methods using bone SPECT/CT with those using planar bone scans to assess active sacroiliitis. Methods: We retrospectively reviewed whole-body bone scans and pelvic bone SPECT/CTs of 8 patients who had clinically confirmed sacroiliitis and enrolled 24 patients without sacroiliitis as references. The volume of interest of each sacroiliac joint, including both the ilium and sacrum, was drawn. Active arthritis zone (AAZ) was defined as the zone of voxels with higher SUV than sacral mean SUV within the VOI of SI joint. Then, the following SPECT/CT quantitative parameters, SUVmax (maximum SUV), SUV50% (mean SUV in highest 50% of SUV), and SUV-AAZ, and the ratio of those values to sacral mean SUV (SUVmax/S, SUV50%/S, SUV-AAZ/S) were calculated. For the planar bone scan, the mean count ratio of SI joint/sacrum (SI/S) was conventionally measured. Results: Most of the SPECT/CT parameters of the sacroiliitis group were significantly higher than the normal group, whereas SI/S of the planar bone scan was not significantly different between the two groups. In receiver operating characteristic curve analysis, SUV-AAZ/S showed the highest AUC of 0.992, followed by SUV50%/S and SUVmax/S. All ratio parameters of the SPECT/CT showed higher AUC values than the SUV parameters of SI joint or SI/S of the planar scan. Conclusions: The quantitative analyses of bone SPECT/CT showed better performance in assessing active sacroiliitis than the planar bone scan. SPECT/CT parameters using the ratio of the SI joint to sacrum showed more favorable results than SUV parameters such as SUVmax, SUV50%, and SUV-AAZ.

The impact of subthreshold levels of amyloid deposition on conversion to dementia in patients with amyloid-negative amnestic mild cognitive impairment.

BACKGROUND: About 40-50% of patients with amnestic mild cognitive impairment (MCI) are found to have no significant Alzheimer's pathology based on amyloid PET positivity. Notably, conversion to dementia in this population is known to occur much less often than in amyloid-positive MCI. However, the relationship between MCI and brain amyloid deposition remains largely unknown. Therefore, we investigated the influence of subthreshold levels of amyloid deposition on conversion to dementia in amnestic MCI patients with negative amyloid PET scans. METHODS: This study was a retrospective cohort study of patients with amyloid-negative amnestic MCI who visited the memory clinic of Asan Medical Center. All participants underwent detailed neuropsychological testing, brain magnetic resonance imaging, and [18F]-florbetaben (FBB) positron emission tomography scan (PET). Conversion to dementia was determined by a neurologist based on a clinical interview with a detailed neuropsychological test or a decline in the Korean version of the Mini-Mental State Examination score of more than 4 points per year combined with impaired activities of daily living. Regional cortical amyloid levels were calculated, and a receiver operating characteristic (ROC) curve for conversion to dementia was obtained. To increase the reliability of the results of the study, we analyzed the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset together. RESULTS: During the follow-up period, 36% (39/107) of patients converted to dementia from amnestic MCI. The dementia converter group displayed increased standardized uptake value ratio (SUVR) values of FBB on PET in the bilateral temporal, parietal, posterior cingulate, occipital, and left precuneus cortices as well as increased global SUVR. Among volume of interests, the left parietal SUVR predicted conversion to dementia with the highest accuracy in the ROC analysis (area under the curve [AUC] = 0.762, P < 0.001). The combination of precuneus, parietal cortex, and FBB composite SUVRs also showed a higher accuracy in predicting conversion to dementia than other models (AUC = 0.763). Of the results of ADNI data, the SUVR of the left precuneus SUVR showed the highest AUC (AUC = 0.596, P = 0.006). CONCLUSION: Our findings suggest that subthreshold amyloid levels may contribute to conversion to dementia in patients with amyloid-negative amnestic MCI.

[18F]THK-5351 PET Patterns in Patients With Alzheimer's Disease and Negative Amyloid PET Findings.

BACKGROUND AND PURPOSE: Alzheimer's disease (AD) does not always mean amyloid positivity. [18F]THK-5351 has been shown to be able to detect reactive astrogliosis as well as tau accompanied by neurodegenerative changes. We evaluated the [18F]THK-5351 retention patterns in positron-emission tomography (PET) and the clinical characteristics of patients clinically diagnosed with AD dementia who had negative amyloid PET findings. METHODS: We performed 3.0-T magnetic resonance imaging, [18F]THK-5351 PET, and amyloid PET in 164 patients with AD dementia. Amyloid PET was visually scored as positive or negative. [18F]THK-5351 PET were visually classified as having an intratemporal or extratemporal spread pattern. RESULTS: The 164 patients included 23 (14.0%) who were amyloid-negative (age 74.9±8.3 years, mean±standard deviation; 9 males, 14 females). Amyloid-negative patients were older, had a higher prevalence of diabetes mellitus, and had better visuospatial and memory functions. The frequency of the apolipoprotein E ε4 allele was higher and the hippocampal volume was smaller in amyloid-positive patients. [18F]THK-5351 uptake patterns of the amyloid-negative patients were classified into intratemporal spread (n=10) and extratemporal spread (n=13). Neuropsychological test results did not differ significantly between these two groups. The standardized uptake value ratio of [18F]THK-5351 was higher in the extratemporal spread group (2.01±0.26 vs. 1.61±0.15, p=0.001). After 1 year, Mini Mental State Examination (MMSE) scores decreased significantly in the extratemporal spread group (-3.5±3.2, p=0.006) but not in the intratemporal spread group (-0.5±2.8, p=0.916). The diagnosis remained as AD (n=5, 50%) or changed to other diagnoses (n=5, 50%) in the intratemporal group, whereas it remained as AD (n=8, 61.5%) or changed to frontotemporal dementia (n=4, 30.8%) and other diagnoses (n=1, 7.7%) in the extratemporal spread group. CONCLUSIONS: Approximately 70% of the patients with amyloid-negative AD showed abnormal [18F]THK-5351 retention. MMSE scores deteriorated rapidly in the patients with an extratemporal spread pattern.

Fully Automatic Quantitative Measurement of 18F-FDG PET/CT in Thymic Epithelial Tumors Using a Convolutional Neural Network.

OBJECTIVES: The aim of this study was to develop a deep learning (DL)-based segmentation algorithm for automatic measurement of metabolic parameters of 18F-FDG PET/CT in thymic epithelial tumors (TETs), comparable performance to manual volumes of interest. PATIENTS AND METHODS: A total of 186 consecutive patients with resectable TETs and preoperative 18F-FDG PET/CT were retrospectively enrolled (145 thymomas, 41 thymic carcinomas). A quasi-3D U-net architecture was trained to resemble ground-truth volumes of interest. Segmentation performance was assessed using the Dice similarity coefficient. Agreements between manual and DL-based automated extraction of SUVmax, metabolic tumor volume (MTV), total lesion glycolysis (TLG), and 63 radiomics features were evaluated via concordance correlation coefficients (CCCs) and linear regression slopes. Diagnostic and prognostic values were compared in terms of area under the receiver operating characteristics curve (AUC) for thymic carcinoma and hazards ratios (HRs) for freedom from recurrence. RESULTS: The mean Dice similarity coefficient was 0.83 ± 0.34. Automatically measured SUVmax (slope, 0.97; CCC, 0.92), MTV (slope, 0.94; CCC, 0.96), and TLG (slope, 0.96; CCC, 0.96) were in good agreement with manual measurements. The mean CCC and slopes were 0.88 ± 0.06 and 0.89 ± 0.05, respectively, for the radiomics parameters. Automatically measured SUVmax, MTV, and TLG showed good diagnostic accuracy for thymic carcinoma (AUCs: SUVmax, 0.95; MTV, 0.85; TLG, 0.87) and significant prognostic value (HRs: SUVmax, 1.31 [95% confidence interval, 1.16-1.48]; MTV, 2.11 [1.09-4.06]; TLG, 1.90 [1.12-3.23]). No significant differences in the AUCs or HRs were found between automatic and manual measurements for any of the metabolic parameters. CONCLUSIONS: Our DL-based model provides comparable segmentation performance and metabolic parameter values to manual measurements in TETs.

Association between tumor 18F-fluorodeoxyglucose metabolism and survival in women with estrogen receptor-positive, HER2-negative breast cancer.

We examined whether 18F-fluorodeoxyglucose metabolism is associated with distant relapse-free survival (DRFS) and overall survival (OS) in women with estrogen receptor (ER)-positive, HER2-negative breast cancer. This was a cohort study examining the risk factors for survival that had occurred at the start of the study. A cohort from Asan Medical Center, Korea, recruited between November 2007 and December 2014, was included. Patients received anthracycline-based neoadjuvant chemotherapy. The maximum standardized uptake value (SUV) of 18F-fluorodeoxyglucose positron emission tomography/computed tomography (PET/CT) was measured. The analysis included 466 women. The median (interquartile range) follow-up period without distant metastasis or death was 6.2 (5.3-7.6) years. Multivariable analysis of hazard ratio (95% confidence interval [CI]) showed that the middle and high tertiles of SUV were prognostic for DRFS (2.93, 95% CI 1.62-5.30; P < 0.001) and OS (4.87, 95% CI 1.94-12.26; P < 0.001). The 8-year DRFS rates were 90.7% (95% CI 85.5-96.1%) for those in the low tertile of maximum SUV vs. 73.7% (95% CI 68.0-79.8%) for those in the middle and high tertiles of maximum SUV. 18F-fluorodeoxyglucose PET/CT may assess the risk of distant metastasis and death in ER-positive, HER2-negative patients.

Unified Deep Learning-Based Mouse Brain MR Segmentation: Template-Based Individual Brain Positron Emission Tomography Volumes-of-Interest Generation Without Spatial Normalization in Mouse Alzheimer Model.

Although skull-stripping and brain region segmentation are essential for precise quantitative analysis of positron emission tomography (PET) of mouse brains, deep learning (DL)-based unified solutions, particularly for spatial normalization (SN), have posed a challenging problem in DL-based image processing. In this study, we propose an approach based on DL to resolve these issues. We generated both skull-stripping masks and individual brain-specific volumes-of-interest (VOIs-cortex, hippocampus, striatum, thalamus, and cerebellum) based on inverse spatial normalization (iSN) and deep convolutional neural network (deep CNN) models. We applied the proposed methods to mutated amyloid precursor protein and presenilin-1 mouse model of Alzheimer's disease. Eighteen mice underwent T2-weighted MRI and 18F FDG PET scans two times, before and after the administration of human immunoglobulin or antibody-based treatments. For training the CNN, manually traced brain masks and iSN-based target VOIs were used as the label. We compared our CNN-based VOIs with conventional (template-based) VOIs in terms of the correlation of standardized uptake value ratio (SUVR) by both methods and two-sample t-tests of SUVR % changes in target VOIs before and after treatment. Our deep CNN-based method successfully generated brain parenchyma mask and target VOIs, which shows no significant difference from conventional VOI methods in SUVR correlation analysis, thus establishing methods of template-based VOI without SN.

Role of White Matter Abnormalities in the Relationship Between Microbleed Burden and Cognitive Impairment in Cerebral Amyloid Angiopathy.

BACKGROUND: Cerebral amyloid angiopathy (CAA) often presents as cognitive impairment, but the mechanism of cognitive decline is unclear. Recent studies showed that number of microbleeds were associated with cognitive decline. OBJECTIVE: We aimed to investigate how microbleeds contribute to cognitive impairment in association with white matter tract abnormalities or cortical thickness in CAA. METHODS: This retrospective comparative study involved patients with probable CAA according to the Boston criteria (Aß+ CAA) and patients with Alzheimer's disease (Aß+ AD), all of whom showed severe amyloid deposition on amyloid PET. Using mediation analysis, we investigated how FA or cortical thickness mediates the correlation between the number of lobar microbleeds and cognition. RESULTS: We analyzed 30 patients with Aß+ CAA (age 72.2±7.6, female 53.3%) and 30 patients with Aß+ AD (age 71.5±7.6, female 53.3%). The two groups showed similar degrees of cortical amyloid deposition in AD-related regions. The Aß+ CAA group had significantly lower FA values in the clusters of the posterior area than did the Aß+ AD group(family-wise error-corrected p < 0.05). The correlation between the number of lobar microbleeds and visuospatial function was indirectly mediated by white matter tract abnormality of right posterior thalamic radiation (PTR) and tapetum, while lobar microbleeds and language function was indirectly mediated by the abnormality of left PTR and sagittal stratum. Cortical thickness did not mediate the association between lobar microbleeds and cognition. CONCLUSION: This result supports the hypothesis that microbleeds burden leads to white matter tract damage and subsequent cognitive decline in CAA.

18F-THK5351 PET Positivity and Longitudinal Changes in Cognitive Function in ß-Amyloid-Negative Amnestic Mild Cognitive Impairment.

PURPOSE: Neuroinflammation is considered an important pathway associated with several diseases that result in cognitive decline. 18F-THK5351 positron emission tomography (PET) signals might indicate the presence of neuroinflammation, as well as Alzheimer's disease-type tau aggregates. ß-amyloid (Aß)-negative (Aß-) amnestic mild cognitive impairment (aMCI) may be associated with non-Alzheimer's disease pathophysiology. Accordingly, we investigated associations between 18F-THK5351 PET positivity and cognitive decline among Aß- aMCI patients. MATERIALS AND METHODS: The present study included 25 amyloid PET negative aMCI patients who underwent a minimum of two follow-up neuropsychological evaluations, including clinical dementia rating-sum of boxes (CDR-SOB). The patients were classified into two groups: 18F-THK5351-positive and -negative groups. The present study used a linear mixed effects model to estimate the effects of 18F-THK5351 PET positivity on cognitive prognosis among Aß- aMCI patients. RESULTS: Among the 25 Aß- aMCI patients, 10 (40.0%) were 18F-THK5351 positive. The patients in the 18F-THK5351-positive group were older than those in the 18F-THK5351-negative group (77.4±2.2 years vs. 70.0±5.5 years; p<0.001). There was no difference between the two groups with regard to the proportion of apolipoprotein E ε4 carriers. Interestingly, however, the CDR-SOB scores of the 18F-THK5351-positive group deteriorated at a faster rate than those of the 18F-THK5351-negative group (B=0.003, p=0.033). CONCLUSION: The results of the present study suggest that increased 18F-THK5351 uptake might be a useful predictor of poor prognosis among Aß- aMCI patients, which might be associated with increased neuroinflammation (ClinicalTrials.gov NCT02656498).

Diagnostic performance of deep learning models for detecting bone metastasis on whole-body bone scan in prostate cancer.

PURPOSE: We evaluated the performance of deep learning classifiers for bone scans of prostate cancer patients. METHODS: A total of 9113 consecutive bone scans (5342 prostate cancer patients) were initially evaluated. Bone scans were labeled as positive/negative for bone metastasis using clinical reports and image review for ground truth diagnosis. Two different 2D convolutional neural network (CNN) architectures were proposed: (1) whole body-based (WB) and (2) tandem architectures integrating whole body and local patches, here named as "global-local unified emphasis" (GLUE). Both models were trained using abundant (72%:8%:20% for training:validation:test sets) and limited training data (10%:40%:50%). The allocation of test sets was rotated across all images: therefore, fivefold and twofold cross-validation test results were available for abundant and limited settings, respectively. RESULTS: A total of 2991 positive and 6142 negative bone scans were used as input. For the abundant training setting, the receiver operating characteristics curves of both the GLUE and WB models indicated excellent diagnostic ability in terms of the area under the curve (GLUE: 0.936-0.955, WB: 0.933-0.957, P > 0.05 in four of the fivefold tests). The overall accuracies of the GLUE and WB models were 0.900 and 0.889, respectively. With the limited training setting, the GLUE models showed significantly higher AUCs than the WB models (0.894-0.908 vs. 0.870-0.877, P < 0.0001). CONCLUSION: Our 2D-CNN models accurately classified bone scans of prostate cancer patients. While both showed excellent performance with the abundant dataset, the GLUE model showed higher performance than the WB model in the limited data setting.