| 1. |
- Sadik, May, 1970, et al.
(author)
-
Metabolic tumour volume in Hodgkin lymphoma - A comparison between manual and AI-based analysis
- 2024
-
In: Clinical Physiology and Functional Imaging. - 1475-0961 .- 1475-097X. ; 44:3, s. 220-227
-
Journal article (peer-reviewed)abstract
- Aim: To compare total metabolic tumour volume (tMTV), calculated using two artificial intelligence (AI)-based tools, with manual segmentation by specialists as the reference. Methods: Forty-eight consecutive Hodgkin lymphoma (HL) patients staged with [18F] fluorodeoxyglucose positron emission tomography/computed tomography were included. The median age was 35 years (range: 7–75), 46% female. The tMTV was automatically measured using the AI-based tools positron emission tomography assisted reporting system (PARS) (from Siemens) and RECOMIA (recomia.org) without any manual adjustments. A group of eight nuclear medicine specialists manually segmented lesions for tMTV calculations; each patient was independently segmented by two specialists. Results: The median of the manual tMTV was 146 cm3 (interquartile range [IQR]: 79–568 cm3) and the median difference between two tMTV values segmented by different specialists for the same patient was 26 cm3 (IQR: 10–86 cm3). In 22 of the 48 patients, the manual tMTV value was closer to the RECOMIA tMTV value than to the manual tMTV value segmented by the second specialist. In 11 of the remaining 26 patients, the difference between the RECOMIA tMTV and the manual tMTV was small (<26 cm3, which was the median difference between two manual tMTV values from the same patient). The corresponding numbers for PARS were 18 and 10 patients, respectively. Conclusion: The results of this study indicate that RECOMIA and Siemens PARS AI tools could be used without any major manual adjustments in 69% (33/48) and 58% (28/48) of HL patients, respectively. This demonstrates the feasibility of using AI tools to support physicians measuring tMTV for assessment of prognosis in clinical practice.
|
|
| 2. |
|
|
| 3. |
- Borrelli, P., et al.
(author)
-
Automated classification of PET-CT lesions in lung cancer: An independent validation study
- 2022
-
In: Clinical Physiology and Functional Imaging. - : Wiley. - 1475-0961 .- 1475-097X. ; 42:5, s. 327-332
-
Journal article (peer-reviewed)abstract
- Introduction Recently, a tool called the positron emission tomography (PET)-assisted reporting system (PARS) was developed and presented to classify lesions in PET/computed tomography (CT) studies in patients with lung cancer or lymphoma. The aim of this study was to validate PARS with an independent group of lung-cancer patients using manual lesion segmentations as a reference standard, as well as to evaluate the association between PARS-based measurements and overall survival (OS). Methods This study retrospectively included 115 patients who had undergone clinically indicated (18F)-fluorodeoxyglucose (FDG) PET/CT due to suspected or known lung cancer. The patients had a median age of 66 years (interquartile range [IQR]: 61-72 years). Segmentations were made manually by visual inspection in a consensus reading by two nuclear medicine specialists and used as a reference. The research prototype PARS was used to automatically analyse all the PET/CT studies. The PET foci classified as suspicious by PARS were compared with the manual segmentations. No manual corrections were applied. Total lesion glycolysis (TLG) was calculated based on the manual and PARS-based lung-tumour segmentations. Associations between TLG and OS were investigated using Cox analysis. Results PARS showed sensitivities for lung tumours of 55.6% per lesion and 80.2% per patient. Both manual and PARS TLG were significantly associated with OS. Conclusion Automatically calculated TLG by PARS contains prognostic information comparable to manually measured TLG in patients with known or suspected lung cancer. The low sensitivity at both the lesion and patient levels makes the present version of PARS less useful to support clinical reading, reporting and staging.
|
|
| 4. |
- Borrelli, P., et al.
(author)
-
Freely available convolutional neural network-based quantification of PET/CT lesions is associated with survival in patients with lung cancer
- 2022
-
In: EJNMMI Physics. - : Springer Science and Business Media LLC. - 2197-7364. ; 9:1
-
Journal article (peer-reviewed)abstract
- Background: Metabolic positron emission tomography/computed tomography (PET/CT) parameters describing tumour activity contain valuable prognostic information, but to perform the measurements manually leads to both intra- and inter-reader variability and is too time-consuming in clinical practice. The use of modern artificial intelligence-based methods offers new possibilities for automated and objective image analysis of PET/CT data. Purpose: We aimed to train a convolutional neural network (CNN) to segment and quantify tumour burden in [18F]-fluorodeoxyglucose (FDG) PET/CT images and to evaluate the association between CNN-based measurements and overall survival (OS) in patients with lung cancer. A secondary aim was to make the method available to other researchers. Methods: A total of 320 consecutive patients referred for FDG PET/CT due to suspected lung cancer were retrospectively selected for this study. Two nuclear medicine specialists manually segmented abnormal FDG uptake in all of the PET/CT studies. One-third of the patients were assigned to a test group. Survival data were collected for this group. The CNN was trained to segment lung tumours and thoracic lymph nodes. Total lesion glycolysis (TLG) was calculated from the CNN-based and manual segmentations. Associations between TLG and OS were investigated using a univariate Cox proportional hazards regression model. Results: The test group comprised 106 patients (median age, 76years (IQR 61–79); n = 59 female). Both CNN-based TLG (hazard ratio 1.64, 95% confidence interval 1.21–2.21; p = 0.001) and manual TLG (hazard ratio 1.54, 95% confidence interval 1.14–2.07; p = 0.004) estimations were significantly associated with OS. Conclusion: Fully automated CNN-based TLG measurements of PET/CT data showed were significantly associated with OS in patients with lung cancer. This type of measurement may be of value for the management of future patients with lung cancer. The CNN is publicly available for research purposes. © 2022, The Author(s).
|
|
| 5. |
|
|
| 6. |
|
|
| 7. |
- Sachpekidis, C., et al.
(author)
-
Application of an artificial intelligence-based tool in F-18 FDG PET/CT for the assessment of bone marrow involvement in multiple myeloma
- 2023
-
In: European Journal of Nuclear Medicine and Molecular Imaging. - 1619-7070 .- 1619-7089. ; 50:12, s. 3697-3708
-
Journal article (peer-reviewed)abstract
- Purpose[F-18]FDG PET/CT is an imaging modality of high performance in multiple myeloma (MM). Nevertheless, the inter-observer reproducibility in PET/CT scan interpretation may be hampered by the different patterns of bone marrow (BM) infiltration in the disease. Although many approaches have been recently developed to address the issue of standardization, none can yet be considered a standard method in the interpretation of PET/CT. We herein aim to validate a novel three-dimensional deep learning-based tool on PET/CT images for automated assessment of the intensity of BM metabolism in MM patients.Materials and methodsWhole-body [F-18]FDG PET/CT scans of 35 consecutive, previously untreated MM patients were studied. All patients were investigated in the context of an open-label, multicenter, randomized, active-controlled, phase 3 trial (GMMG-HD7). Qualitative (visual) analysis classified the PET/CT scans into three groups based on the presence and number of focal [F-18]FDG-avid lesions as well as the degree of diffuse [F-18]FDG uptake in the BM. The proposed automated method for BM metabolism assessment is based on an initial CT-based segmentation of the skeleton, its transfer to the SUV PET images, the subsequent application of different SUV thresholds, and refinement of the resulting regions using postprocessing. In the present analysis, six different SUV thresholds (Approaches 1-6) were applied for the definition of pathological tracer uptake in the skeleton [Approach 1: liver SUVmedian x 1.1 (axial skeleton), gluteal muscles SUVmedian x 4 (extremities). Approach 2: liver SUVmedian x 1.5 (axial skeleton), gluteal muscles SUVmedian x 4 (extremities). Approach 3: liver SUVmedian x 2 (axial skeleton), gluteal muscles SUVmedian x 4 (extremities). Approach 4: & GE; 2.5. Approach 5: & GE; 2.5 (axial skeleton), & GE; 2.0 (extremities). Approach 6: SUVmax liver]. Using the resulting masks, subsequent calculations of the whole-body metabolic tumor volume (MTV) and total lesion glycolysis (TLG) in each patient were performed. A correlation analysis was performed between the automated PET values and the results of the visual PET/CT analysis as well as the histopathological, cytogenetical, and clinical data of the patients.ResultsBM segmentation and calculation of MTV and TLG after the application of the deep learning tool were feasible in all patients. A significant positive correlation (p < 0.05) was observed between the results of the visual analysis of the PET/CT scans for the three patient groups and the MTV and TLG values after the employment of all six [F-18]FDG uptake thresholds. In addition, there were significant differences between the three patient groups with regard to their MTV and TLG values for all applied thresholds of pathological tracer uptake. Furthermore, we could demonstrate a significant, moderate, positive correlation of BM plasma cell infiltration and plasma levels of & beta;2-microglobulin with the automated quantitative PET/CT parameters MTV and TLG after utilization of Approaches 1, 2, 4, and 5.ConclusionsThe automated, volumetric, whole-body PET/CT assessment of the BM metabolic activity in MM is feasible with the herein applied method and correlates with clinically relevant parameters in the disease. This methodology offers a potentially reliable tool in the direction of optimization and standardization of PET/CT interpretation in MM. Based on the present promising findings, the deep learning-based approach will be further evaluated in future prospective studies with larger patient cohorts.
|
|
| 8. |
|
|
| 9. |
- Abuhasanein, Suleiman, et al.
(author)
-
A novel model of artificial intelligence based automated image analysis of CT urography to identify bladder cancer in patients investigated for macroscopic hematuria
- 2024
-
In: Scandinavian Journal of Urology. - : Medical Journal Sweden AB. - 2168-1805 .- 2168-1813. ; 59, s. 90-97
-
Journal article (peer-reviewed)abstract
- OBJECTIVE: To evaluate whether artificial intelligence (AI) based automatic image analysis utilising convolutional neural networks (CNNs) can be used to evaluate computed tomography urography (CTU) for the presence of urinary bladder cancer (UBC) in patients with macroscopic hematuria. METHODS: Our study included patients who had undergone evaluation for macroscopic hematuria. A CNN-based AI model was trained and validated on the CTUs included in the study on a dedicated research platform (Recomia.org). Sensitivity and specificity were calculated to assess the performance of the AI model. Cystoscopy findings were used as the reference method. RESULTS: The training cohort comprised a total of 530 patients. Following the optimisation process, we developed the last version of our AI model. Subsequently, we utilised the model in the validation cohort which included an additional 400 patients (including 239 patients with UBC). The AI model had a sensitivity of 0.83 (95% confidence intervals [CI], 0.76-0.89), specificity of 0.76 (95% CI 0.67-0.84), and a negative predictive value (NPV) of 0.97 (95% CI 0.95-0.98). The majority of tumours in the false negative group (n = 24) were solitary (67%) and smaller than 1 cm (50%), with the majority of patients having cTaG1-2 (71%). CONCLUSIONS: We developed and tested an AI model for automatic image analysis of CTUs to detect UBC in patients with macroscopic hematuria. This model showed promising results with a high detection rate and excessive NPV. Further developments could lead to a decreased need for invasive investigations and prioritising patients with serious tumours.
|
|
| 10. |
- Berg, J, et al.
(author)
-
CAVIAR : a tool to improve serial analysis of the 12-lead electrocardiogram
- 1995
-
In: Clinical Physiology and Functional Imaging. - : Wiley. - 0144-5979. ; 15:5, s. 435-445
-
Journal article (peer-reviewed)abstract
- An important part of an electrocardiogram (ECG) interpretation is the comparison between the present ECG and earlier recordings. The purpose of the present study was to evaluate a combination of two computer-based methods, synthesized vectorcardiogram (VCG) and CAVIAR, in this comparison. The methods were applied to a group of 38 normal subjects and to a group of 36 patients treated with anthracyclines. A fraction of these patients are likely to develop cardiac injury during or after the treatment, since anthracyclines are known to cause heart failure and cardiomyopathy. Two ECGs were recorded on each patient, one before and one after the treatment. On each normal subject, two ECGs were recorded with an interval of 8-9 years. A synthesized VCG was calculated from each ECG and the two synthesized VCGs from each subject were analysed with the CAVIAR method. The CAVIAR analysis is a quantitative method and normal limits for four measurements were established using the normal group. Values above these limits were more frequent in the patient group than in the normal group. The conventional ECGs were also analysed visually by an experience ECG interpreter without knowledge of the result of the CAVIAR analysis. No significant serial changes were found in 10 of the patients with high CAVIAR values. Changes in the ECGs were found in two patients with normal CAVIAR values. In summary, synthesized VCG and CAVIAR could be used to highlight small serial changes that are difficult to find in a visual analysis of ECGs.
|
|
| 11. |
|
|
| 12. |
|
|
| 13. |
|
|
| 14. |
- Forberg, Jakob L, et al.
(author)
-
An artificial neural network to safely reduce the number of ambulance ECGs transmitted for physician assessment in a system with prehospital detection of ST elevation myocardial infarction
- 2012
-
In: Scandinavian Journal of Trauma, Resuscitation and Emergency Medicine. - : Springer Science and Business Media LLC. - 1757-7241. ; 20:1, s. 1-9
-
Journal article (peer-reviewed)abstract
- Background: Pre-hospital electrocardiogram (ECG) transmission to an expert for interpretation and triage reduces time to acute percutaneous coronary intervention (PCI) in patients with ST elevation Myocardial Infarction (STEMI). In order to detect all STEMI patients, the ECG should be transmitted in all cases of suspected acute cardiac ischemia. The aim of this study was to examine the ability of an artificial neural network (ANN) to safely reduce the number of ECGs transmitted by identifying patients without STEMI and patients not needing acute PCI. Methods: Five hundred and sixty ambulance ECGs transmitted to the coronary care unit (CCU) in routine care were prospectively collected. The ECG interpretation by the ANN was compared with the diagnosis (STEMI or not) and the need for an acute PCI (or not) as determined from the Swedish coronary angiography and angioplasty register. The CCU physician's real time ECG interpretation (STEMI or not) and triage decision (acute PCI or not) were registered for comparison. Results: The ANN sensitivity, specificity, positive and negative predictive values for STEMI was 95%, 68%, 18% and 99%, respectively, and for a need of acute PCI it was 97%, 68%, 17% and 100%. The area under the ANN's receiver operating characteristics curve for STEMI detection was 0.93 (95% CI 0.89-0.96) and for predicting the need of acute PCI 0.94 (95% CI 0.90-0.97). If ECGs where the ANN did not identify a STEMI or a need of acute PCI were theoretically to be withheld from transmission, the number of ECGs sent to the CCU could have been reduced by 64% without missing any case with STEMI or a need of immediate PCI. Conclusions: Our ANN had an excellent ability to predict STEMI and the need of acute PCI in ambulance ECGs, and has a potential to safely reduce the number of ECG transmitted to the CCU by almost two thirds.
|
|
| 15. |
- Forberg, Jakob L, et al.
(author)
-
Direct hospital costs of chest pain patients attending the emergency department: a retrospective study
- 2006
-
In: BMC Emergency Medicine. - 1471-227X. ; 6
-
Journal article (peer-reviewed)abstract
- BACKGROUND: Chest pain is one of the most common complaints in the Emergency Department (ED), but the cost of ED chest pain patients is unclear. The aim of this study was to describe the direct hospital costs for unselected chest pain patients attending the emergency department (ED). METHODS: 1,000 consecutive ED visits of patients with chest pain were retrospectively included. Costs directly following the ED visit were retrieved from the hospital economy system. RESULTS: The mean cost per patient visit was 26.8 thousand Swedish kronar (kSEK) (median 7.2 kSEK), with admission time accounting for 73% of all costs. Mean cost for patients discharged from the ED was 1.4 kSEK (median 1.3 kSEK), and for patients without ACS admitted 1 day or less 7.6 kSEK (median 6.9 kSEK). The practice in the present study to admit 67% of the patients, of whom only 31% proved to have ACS, was estimated to give a cost per additional life-year saved by hospital admission, compared to theoretical strategy of discharging all patients home, of about 350 kSEK (39 kEUR or 42 kUSD). CONCLUSION: Costs for chest pain patients are large and primarily due to admission time. The present admission practice seems to be cost-effective, but the substantial overadmission indicates that better ED diagnostics and triage could decrease costs considerably.
|
|
| 16. |
- Forberg, Jakob L, et al.
(author)
-
In search of the best method to predict acute coronary syndrome using only the electrocardiogram from the emergency department.
- 2009
-
In: Journal of Electrocardiology. - : Elsevier BV. - 1532-8430 .- 0022-0736. ; 42:1, s. 58-63
-
Journal article (peer-reviewed)abstract
- INTRODUCTION: The aim of this study was to compare different methods to predict acute coronary syndrome (ACS) using only data from a single electrocardiogram (ECG) in the emergency department (ED). METHOD: We compared the ACS prediction abilities of classical ECG criteria, human expert ECG interpretation, a logistic regression model and an artificial neural network ensemble (ANN). The ED ECG and discharge diagnoses were retrieved for 861 patient visits to the ED for chest pain. Cross-validation was used to estimate the generalization performance of the logistic regression and the ANN model. RESULTS: The logistic regression model had the overall best performance in predicting ACS with an area under the receiver operating characteristic curve of 0.88. The sensitivities of logistic regression, ANN, expert physicians, and classical ECG criteria were 95%, 95%, 82%, and 75%, respectively, and the specificities were 54%, 44%, 63%, and 69%. CONCLUSION: Our logistic regression model was the best overall method to predict ACS, followed by our ANN. Decision support models have the potential to improve even experienced ECG readers' ability to predict ACS in the ED.
|
|
| 17. |
- Holst, H., et al.
(author)
-
Intelligent computer reporting 'lack of experience' : A confidence measure for decision support systems
- 1998
-
In: Clinical Physiology. - : Wiley. - 0144-5979. ; 18:2, s. 139-147
-
Journal article (peer-reviewed)abstract
- The purpose of this study was to explore the feasibility of developing artificial neural networks that are able to provide confidence measures for their diagnostic advice. Computer-aided decision making can improve physician performance, but many physicians hesitate to use these 'black boxes'. If we are to rely upon decision support systems for such tasks as medical diagnosis it is essential that the computers indicate when the advice given is based on experience, i.e. give a confidence measure. An artificial neural network was trained to diagnose healed anterior myocardial infarction and to indicate 'lack of experience' when test electrocardiograms were different from the electrocardiograms of the training set. A database of 1249 electrocardiograms from patients who had undergone cardiac catheterization was used to train and test the neural network. Thereafter, the ability of the network to indicate 'lack of experience' was assessed using 100 left bundle branch block electrocardiograms, an electrocardiographic pattern that was excluded from the training set. The network indicated that 83% of the left bundle branch block electrocardiograms and 1% of the test electrocardiograms from catheterized patients were different from the electrocardiograms of the training set. All but one of the left bundle branch block electrocardiograms would otherwise be falsely classified as anterior myocardial infarction by the network. Artificial neural networks can be trained to indicate 'lack of experience', and this ability increases the possibility for neural networks to be accepted as reliable decision support systems in clinical practice.
|
|
| 18. |
|
|
| 19. |
- Lindgren Belal, Sarah, et al.
(author)
-
3D skeletal uptake of F-18 sodium fluoride in PET/CT images is associated with overall survival in patients with prostate cancer
- 2017
-
In: EJNMMI Research. - : Springer Science and Business Media LLC. - 2191-219X. ; 7:1
-
Journal article (peer-reviewed)abstract
- Background: Sodium fluoride (NaF) positron emission tomography combined with computer tomography (PET/CT) has shown to be more sensitive than the whole-body bone scan in the detection of skeletal uptake due to metastases in prostate cancer. We aimed to calculate a 3D index for NaF PET/CT and investigate its correlation to the bone scan index (BSI) and overall survival (OS) in a group of patients with prostate cancer. Methods: NaF PET/CT and bone scans were studied in 48 patients with prostate cancer. Automated segmentation of the thoracic and lumbar spines, sacrum, pelvis, ribs, scapulae, clavicles, and sternum were made in the CT images. Hotspots in the PET images were selected using both a manual and an automated method. The volume of each hotspot localized in the skeleton in the corresponding CT image was calculated. Two PET/CT indices, based on manual (manual PET index) and automatic segmenting using a threshold of SUV 15 (automated PET15 index), were calculated by dividing the sum of all hotspot volumes with the volume of all segmented bones. BSI values were obtained using a software for automated calculations. Results: BSI, manual PET index, and automated PET15 index were all significantly associated with OS and concordance indices were 0.68, 0.69, and 0.70, respectively. The median BSI was 0.39 and patients with a BSI > 0.39 had a significantly shorter median survival time than patients with a BSI 0.53 had a significantly shorter median survival time than patients with a manual PET index 0.11 had a significantly shorter median survival time than patients with an automated PET15 index
|
|
| 20. |
- Lomsky, Milan, 1948, et al.
(author)
-
A new automated method for analysis of gated-SPECT images based on a three-dimensional heart shaped model
- 2005
-
In: Clin Physiol Funct Imaging. - 1475-0961. ; 25:4, s. 234-40
-
Journal article (peer-reviewed)abstract
- A new automated method for quantification of left ventricular function from gated-single photon emission computed tomography (SPECT) images has been developed. The method for quantification of cardiac function (CAFU) is based on a heart shaped model and the active shape algorithm. The model contains statistical information of the variability of left ventricular shape. CAFU was adjusted based on the results from the analysis of five simulated gated-SPECT studies with well defined volumes of the left ventricle. The digital phantom NURBS-based Cardiac-Torso (NCAT) and the Monte-Carlo method SIMIND were used to simulate the studies. Finally CAFU was validated on ten rest studies from patients referred for routine stress/rest myocardial perfusion scintigraphy and compared with Cedar-Sinai quantitative gated-SPECT (QGS), a commercially available program for quantification of gated-SPECT images. The maximal differences between the CAFU estimations and the true left ventricular volumes of the digital phantoms were 11 ml for the end-diastolic volume (EDV), 3 ml for the end-systolic volume (ESV) and 3% for the ejection fraction (EF). The largest differences were seen in the smallest heart. In the patient group the EDV calculated using QGS and CAFU showed good agreement for large hearts and higher CAFU values compared with QGS for the smaller hearts. In the larger hearts, ESV was much larger for QGS than for CAFU both in the phantom and patient studies. In the smallest hearts there was good agreement between QGS and CAFU. The findings of this study indicate that our new automated method for quantification of gated-SPECT images can accurately measure left ventricular volumes and EF.
|
|
| 21. |
|
|
| 22. |
- Lomsky, Milan, 1948, et al.
(author)
-
Validation of a new automated method for analysis of gated-SPECT images
- 2006
-
In: Clin Physiol Funct Imaging. - 1475-0961. ; 26:3, s. 139-45
-
Journal article (peer-reviewed)abstract
- We recently presented a new method for quantification of CArdiac FUnction--denoted CAFU--as the first step in the development of an automated method for integrated interpretation of gated myocardial perfusion single photon emission computed tomography (SPECT) images. The aim of this study was to validate CAFU in the assessment of global and regional function of the left ventricle. Quantitative gated-SPECT (QGS), the most widely used software package for quantification of gated-SPECT images, was used as reference method for the measurements of ejection fraction (EF) and ventricular volumes, and visual analysis by an experienced physician was used as reference method for the measurements of regional wall motion and thickening. Two different groups of consecutive patients referred for myocardial perfusion scintigraphy were studied. Global function was evaluated in 316 patients and regional function in 49 other patients. The studies were performed using a 2-day stress/rest 99 m-Tc-sestamibi protocol. A good correlation was found between EF values from QGS and CAFU (EF CAFU = 0.84 EF QGS + 13, r = 0.94), but CAFU values were on average 4 EF points higher than QGS values. With CAFU the segments with normal thickening according to the physician showed significantly higher thickening values (in all parts of the myocardium) compared to the segments classified as having abnormal thickening. In conclusion, this study demonstrates that CAFU can be used to quantify global and regional function in gated-SPECT images. This is an important step in our development of an automated method for integrated interpretation of gated-SPECT myocardial perfusion scintigraphy studies.
|
|
| 23. |
|
|
| 24. |
|
|
| 25. |
- Saito, Shintaro, et al.
(author)
-
Convolutional neural network-based automatic heart segmentation and quantitation in 123 I-metaiodobenzylguanidine SPECT imaging
- 2021
-
In: EJNMMI Research. - : Springer Science and Business Media LLC. - 2191-219X. ; 11:1
-
Journal article (peer-reviewed)abstract
- Background: Since three-dimensional segmentation of cardiac region in 123I-metaiodobenzylguanidine (MIBG) study has not been established, this study aimed to achieve organ segmentation using a convolutional neural network (CNN) with 123I-MIBG single photon emission computed tomography (SPECT) imaging, to calculate heart counts and washout rates (WR) automatically and to compare with conventional quantitation based on planar imaging. Methods: We assessed 48 patients (aged 68.4 ± 11.7 years) with heart and neurological diseases, including chronic heart failure, dementia with Lewy bodies, and Parkinson's disease. All patients were assessed by early and late 123I-MIBG planar and SPECT imaging. The CNN was initially trained to individually segment the lungs and liver on early and late SPECT images. The segmentation masks were aligned, and then, the CNN was trained to directly segment the heart, and all models were evaluated using fourfold cross-validation. The CNN-based average heart counts and WR were calculated and compared with those determined using planar parameters. The CNN-based SPECT and conventional planar heart counts were corrected by physical time decay, injected dose of 123I-MIBG, and body weight. We also divided WR into normal and abnormal groups from linear regression lines determined by the relationship between planar WR and CNN-based WR and then analyzed agreement between them. Results: The CNN segmented the cardiac region in patients with normal and reduced uptake. The CNN-based SPECT heart counts significantly correlated with conventional planar heart counts with and without background correction and a planar heart-to-mediastinum ratio (R2 = 0.862, 0.827, and 0.729, p < 0.0001, respectively). The CNN-based and planar WRs also correlated with and without background correction and WR based on heart-to-mediastinum ratios of R2 = 0.584, 0.568 and 0.507, respectively (p < 0.0001). Contingency table findings of high and low WR (cutoffs: 34% and 30% for planar and SPECT studies, respectively) showed 87.2% agreement between CNN-based and planar methods. Conclusions: The CNN could create segmentation from SPECT images, and average heart counts and WR were reliably calculated three-dimensionally, which might be a novel approach to quantifying SPECT images of innervation.
|
|
