Automatic detection of primary and metastatic lesions on cervicothoracic region and whole-body bone using a uniform machine-learnable approach for [18F]-FDG-PET/CT image analysis.

We propose a method to detect primary and metastatic lesions with Fluorine-18 fluorodeoxyglucose (FDG) accumulation in the lung field, neck, mediastinum, and bony regions on the FDG-PET/CT images. To search for systemic lesions, various anatomical structures must be considered. The proposed method is addressed by using an extraction process for anatomical regions and a uniform lesion detection approach. The uniform approach does not utilize processes that reflect any region-specific anatomical aspects but has a machine-learnable framework. Therefore, it can work as a lesion detection process for a specific anatomical region if it machine-learns the specific region data. In this study, three lesion detection processes for the whole-body bone region, lung field, or neck-mediastinum region are obtained. These detection processes include lesion candidate detection and false positive (FP) candidate elimination. The lesion candidate detection is based on a voxel anomaly detection with a one-class support vector machine. The FP candidate elimination is performed using an AdaBoost classifier ensemble. The image features used by the ensemble are selected sequentially during training and are optimal for candidate classification. Three-fold cross-validation was used to detect performance with the 54 diseased FDG-PET/CT images. The mean sensitivity for detecting primary and metastatic lesions at 3 FPs per case was 0.89 with a 0.10 standard deviation (SD) in the bone region, 0.80 with a 0.10 SD in the lung field, and 0.87 with a 0.10 SD in the neck region. The average areas under the ROC curve were 0.887 with a 0.125 SD for detecting bone metastases, 0.900 with a 0.063 SD for detecting pulmonary lesions, and 0.927 with a 0.035 SD for detecting the neck-mediastinum lesions. These detection performances indicate that the proposed method could be applied clinically. These results also show that the uniform approach has high versatility for providing various lesion detection processes.

Utility of U-Net for the objective segmentation of the fibroglandular tissue region on clinical digital mammograms.

This study investigates the equivalence or compatibility between U-Net and visual segmentations of fibroglandular tissue regions by mammography experts for calculating the breast density and mean glandular dose (MGD). A total of 703 mediolateral oblique-view mammograms were used for segmentation. Two region types were set as the ground truth (determined visually): (1) one type included only the region where fibroglandular tissue was identifiable (called the 'dense region'); (2) the other type included the region where the fibroglandular tissue may have existed in the past, provided that apparent adipose-only parts, such as the retromammary space, are excluded (the 'diffuse region'). U-Net was trained to segment the fibroglandular tissue region with an adaptive moment estimation optimiser, five-fold cross-validated with 400 training and 100 validation mammograms, and tested with 203 mammograms. The breast density and MGD were calculated using the van Engeland and Dance formulas, respectively, and compared between U-Net and the ground truth with the Dice similarity coefficient and Bland-Altman analysis. Dice similarity coefficients between U-Net and the ground truth were 0.895 and 0.939 for the dense and diffuse regions, respectively. In the Bland-Altman analysis, no proportional or fixed errors were discovered in either the dense or diffuse region for breast density, whereas a slight proportional error was discovered in both regions for the MGD (the slopes of the regression lines were -0.0299 and -0.0443 for the dense and diffuse regions, respectively). Consequently, the U-Net and ground truth were deemed equivalent (interchangeable) for breast density and compatible (interchangeable following four simple arithmetic operations) for MGD. U-Net-based segmentation of the fibroglandular tissue region was satisfactory for both regions, providing reliable segmentation for breast density and MGD calculations. U-Net will be useful in developing a reliable individualised screening-mammography programme, instead of relying on the visual judgement of mammography experts.

Observation of Skin Color Change by Carotenosis in Hyperlipidemia Patient.

PURPOSE: In this study, we focus on the relationship between carotenosis and hyperlipidemia. Because carotenoids are lipophilic, they are more soluble in blood with high lipid levels. Therefore, it is considered that hyperlipidemic patients show carotenosis more strongly than healthy subjects. METHODS: The palm where carotenosis is likely to appear is measured. Carotenoids are intake by 200 mL of 100% fruit juice of an orange. The average of the red values obtained from the images is evaluated. RESULTS: A significant difference was confirmed from 30 min after the carotenoid intake. CONCLUSION: In the future, increasing the number of subjects and doses of carotenoids will be considered for more reliable verification.

Clinical performance of a novel hyperspectral imaging device for cutaneous melanoma and pigmented skin lesions in Caucasian skin.

BACKGROUND: The quest for diagnostic tools for the detection of cutaneous malignant melanoma (cMM) is ongoing. A challenge in cMM care is not overlooking cMM at an early stage, while simultaneously avoiding unnecessary biopsies or excisions of benign pigmented skin lesions (PSLs). A novel hyperspectral imaging (HSI) device is shown to have potential for differentiating equivocal PSLs in Asian skin types. Our objective was to assess the accuracy of the HSI device in distinguishing between cMM and benign PSLs in patients with Caucasian skin types. METHODS: Patients with Caucasian skin types (Fitzpatrick I-II), enrolled for excisional biopsies of PSLs were included and examined using the HSI device. The discrimination index (DI) was used to demonstrate the sensitivity (SE) and specificity (SP) in comparison with the re-evaluated histopathology diagnoses. RESULTS: In 186 patients, 202 pigmented skin lesions were included. The sensitivity to detect cMM was 96.7% (87/90), and the specificity for benign lesions was 42.1% (45/107). The AUC was 0.800 (95% confidence interval (CI): 0.740-0.861). CONCLUSIONS: Our novel HSI device showed a high sensitivity in detecting malignant lesions in patients with Caucasian skin types. Compared with analogous technologies, as multispectral imaging or electrical impedance spectroscopy, our device showed similar or better accuracy in differentiating cMM from benign PSLs. Therefore, it might be a useful clinical tool in skin types I-IV and where further triage of pigmented skin lesions is important.

Automatic diagnosis of melanoma using hyperspectral data and GoogLeNet.

BACKGROUND: Melanoma is a type of superficial tumor. As advanced melanoma has a poor prognosis, early detection and therapy are essential to reduce melanoma-related deaths. To that end, there is a need to develop a quantitative method for diagnosing melanoma. This paper reports the development of such a diagnostic system using hyperspectral data (HSD) and a convolutional neural network, which is a type of machine learning. MATERIALS AND METHODS: HSD were acquired using a hyperspectral imager, which is a type of spectrometer that can simultaneously capture information about wavelength and position. GoogLeNet pre-trained with Imagenet was used to model the convolutional neural network. As many CNNs (including GoogLeNet) have three input channels, the HSD (involving 84 channels) could not be input directly. For that reason, a "Mini Network" layer was added to reduce the number of channels from 84 to 3 just before the GoogLeNet input layer. In total, 619 lesions (including 278 melanoma lesions and 341 non-melanoma lesions) were used for training and evaluation of the network. RESULTS AND CONCLUSION: The system was evaluated by 5-fold cross-validation, and the results indicate sensitivity, specificity, and accuracy of 69.1%, 75.7%, and 72.7% without data augmentation, 72.3%, 81.2%, and 77.2% with data augmentation, respectively. In future work, it is intended to improve the Mini Network and to increase the number of lesions.

AI approach of cycle-consistent generative adversarial networks to synthesize PET images to train computer-aided diagnosis algorithm for dementia.

OBJECTIVE: An artificial intelligence (AI)-based algorithm typically requires a considerable amount of training data; however, few training images are available for dementia with Lewy bodies and frontotemporal lobar degeneration. Therefore, this study aims to present the potential of cycle-consistent generative adversarial networks (CycleGAN) to obtain enough number of training images for AI-based computer-aided diagnosis (CAD) algorithms for diagnosing dementia. METHODS: We trained CycleGAN using 43 amyloid-negative and 45 positive images in slice-by-slice. RESULTS: The CycleGAN can be used to synthesize reasonable amyloid-positive images, and the continuity of slices was preserved. DISCUSSION: Our results show that CycleGAN has the potential to generate a sufficient number of training images for CAD of dementia.

Automatic delineation algorithm of reference region for amyloid imaging based on kinetics.

OBJECTIVE: This study aims to develop an algorithm named AutoRef to delineate a reference region for quantitative PET amyloid imaging. METHODS: AutoRef sets the reference region automatically using a distinguishing feature in the kinetics of reference region. This is reflected in the shapes of the tissue time activity curve. A statistical shape recognition algorithm of the gaussian mixture model is applied with considering spatial and temporal information on a reference region. We evaluate the BPND with manually set reference region and AutoRef using 86 cases (43 positive cases, 10 equivocal cases, and 33 negative cases) of dynamically scanned 11C-Pittsburgh Compound-B. RESULTS: From the Bland-Altman plot, the difference between two BPND is 0.099 ± 0.21 as standard deviation, and no significant systematic error is observed between the BPND with AutoRef and with manual definition of a reference region. Although a proportional error is detected, it is smaller than the 95% limits of agreement. Therefore, the proportional error is negligibly small. CONCLUSIONS: AutoRef presents the same performance as the manual definition of the reference region. Further, since AutoRef is more algorithmic than the ordinary manual definition of the reference region, there are few operator-oriented uncertainties in AutoRef. We thus conclude that AutoRef can be applied as an automatic delineating algorithm for the reference region in amyloid imaging.

Quantitative cosmetic evaluation of long-lasting foundation using multispectral imaging.

BACKGROUND: We tried to search appropriate wavelength to quantitatively evaluate the ability of long-lasting foundation using a hyperspectral imager (HSI) which can simultaneously measure position and wavelength information. MATERIALS AND METHODS: A good reputable long-lasting foundation was applied to the skin of 10 healthy volunteers. Their skin was measured by our newly developed HSI every 2 hours from immediately after application to 6 hours. The application state of the foundation was quantified using the standard deviation of reflectance. RESULTS: A high correlation between standard deviation and the application state of the foundation was confirmed at many wavelengths. In particular, it was suggested that by using the standard deviation of 800 nm, the application state of the foundation can be evaluated quantitatively without depending on the subject's oxygen saturation level. CONCLUSION: By quantitatively evaluating the cosmetic-applied skin by our system, further efficiency improvement of the volunteer experiment is expected.

Clinical contributions of exhaled volatile organic compounds in the diagnosis of lung cancer.

BACKGROUND: Exhaled volatile organic compounds (VOC) are being considered as biomarkers for various lungs diseases, including cancer. However, the accurate measurement of extremely low concentrations of VOC in expired air is technically challenging. We evaluated the clinical contribution of exhaled VOC measured with a new, double cold-trap method in the diagnosis of lung cancer. METHODS: Breath samples were collected from 116 patients with histologically confirmed lung cancer and 37 healthy volunteers (controls) after inspiration of purified air, synthesized through a cold-trap system. The exhaled VOC, trapped in the same system, were heat extracted. We analyzed 14 VOC with gas chromatography. RESULTS: The concentrations of exhaled cyclohexane and xylene were significantly higher in patients with lung cancer than in controls (p = 0.002 and 0.0001, respectively), increased significantly with the progression of the clinical stage of cancer (both p < 0.001), and decreased significantly after successful treatment of 6 patients with small cell lung cancer (p = 0.06 and 0.03, respectively). CONCLUSION: Measurements of exhaled VOCs by a double cold-trap method may help diagnose lung cancer and monitor its progression and regression.

Editor's Highlight: Prospective Analyses of Volatile Organic Compounds in Obstructive Sleep Apnea Patients.

Various volatile organic compounds (VOCs) are known to be toxic. Although exhaled VOC patterns change in obstructive sleep apnea (OSA) patients, individual VOC profiles are not fully determined. The primary outcome was VOC characterizations; secondary outcomes included their relationships with sleep and clinical parameters in OSA patients. We prospectively examined 32 OSA patients with an apnea-hypopnea index (AHI) ≥ 15 by full polysomnography, and 33 age- and sex-matched controls without obvious OSA symptoms. Nine severe OSA patients were examined before and after continuous positive airway pressure (CPAP) treatment. By applying a method which eliminates environmental VOC influences, exhaled VOCs were identified by gas chromatography (GC)-mass spectrometry, and their concentrations were determined by GC. Exhaled aromatic hydrocarbon concentrations (toluene, ethylbenzene, p-xylene, and phenylacetic acid) in the severe OSA groups (AHI ≥ 30) and exhaled saturated hydrocarbon concentrations (hexane, heptane, octane, nonane, and decane) in the most severe OSA group (AHI ≥ 60) were higher than those in the control group. Exhaled isoprene concentrations were increased in all OSA groups (AHI ≥ 15); acetone concentration was increased in the most severe OSA group. Ethylbenzene, p-xylene, phenylacetic acid, and nonane concentrations were increased according to OSA severity, and correlated with AHI, arousal index, and duration of percutaneous oxygen saturation (SpO2) ≤ 90%. Multiple regression analyses revealed these 4 VOC levels were associated with the duration of SpO2 ≤ 90%. Isoprene and acetone decreased after CPAP treatment. OSA increased some toxic VOCs, and some correlated with OSA severity. CPAP treatment possibly ameliorates these productions.

Portable hyperspectral imager with continuous wave green laser for identification and detection of untreated latent fingerprints on walls.

Untreated latent fingerprints are known to exhibit fluorescence under UV laser excitation. Previously, the hyperspectral imager (HSI) has been primarily evaluated in terms of its potential to enhance the sensitivity of latent fingerprint detection following treatment by conventional chemical methods in the forensic science field. In this study however, the potential usability of the HSI for the visualization and detection of untreated latent fingerprints by measuring their inherent fluorescence under continuous wave (CW) visible laser excitation was examined. Its potential to undertake spectral separation of overlapped fingerprints was also evaluated. The excitation wavelength dependence of fluorescent images was examined using an untreated palm print on a steel based wall, and it was found that green laser excitation is superior to blue and yellow lasers' excitation for the production of high contrast fluorescence images. In addition, a spectral separation method for overlapped fingerprints/palm prints on a plaster wall was proposed using new images converted by the division and subtraction of two single wavelength images constructed based on measured hyperspectral data (HSD). In practical tests, the relative isolation of two overlapped fingerprints/palm prints was successful in twelve out of seventeen cases. Only one fingerprint/palm print was extracted for an additional three cases. These results revealed that the feasibility of overlapped fingerprint/palm print spectral separation depends on the difference in the temporal degeneration of each fluorescence spectrum. The present results demonstrate that a combination of a portable HSI and CW green laser has considerable potential for the identification and detection of untreated latent fingerprints/palm prints on the walls under study, while the use of HSD makes it practically possible for doubly overlapped fingerprints/palm prints to be separated spectrally.

Hyperspectroscopic screening of melanoma on acral volar skin.

BACKGROUND/PURPOSE: Early detection and proper excision of the primary lesions of melanoma are crucial for reducing melanoma-related deaths. To support the early detection of melanoma, automated melanoma-screening systems have been extensively studied and developed. In this article, a previously reported hyperspectral imager and melanoma discrimination index are applied to the discrimination of acral lentiginous melanoma (ALM) from acral nevus (AN), and their diagnostic performance is reported. METHODS: The index expresses the disordered nature of each lesion including variegation in color based on variation in spectral information obtained from each lesion. Performance of the index has been studied in thirteen cases of ALM and seven cases of AN, obtained from patients and volunteers, all of whom were Japanese. RESULTS: The index discriminated ALM from AN with a sensitivity of 92% and a specificity of 86%. The area under the receiver operating characteristic curve was 0.97. CONCLUSION: The performance of the proposed objective melanoma discrimination index at a molecular pigmentary level approached that of clinical experts, using the three-step algorithm as the gold standard.

No accumulation of globotriaosylceramide in the heart of a patient with the E66Q mutation in the α-galactosidase A gene.

BACKGROUND: Fabry disease is an X-linked lysosomal disorder resulting from mutations in the α-galactosidase A (GLA) gene. Recent reports described that the E66Q mutation of GLA is not a disease-causing mutation. However, no pathological study was reported. We carried out pathological studies using a cardiac biopsy specimen from a patient with the E66Q mutation. MATERIALS AND METHODS: The case was a 34 year old male patient with end-stage renal failure and cardiomegaly. He was diagnosed with gout at 15 years of age and hemodialysis was started for gouty nephropathy from 31 years of age. He was suspected of having Fabry disease as the result of a screening study for Fabry disease in patients with end-stage renal failure and was referred to our hospital for mutation analysis of the GLA gene. We carried out enzymatic and genetic analysis for GLA and pathological studies of a cardiac biopsy specimen. RESULTS: The patient had the E66Q mutation in the GLA gene. GLA activity in leukocytes was 36.2% of the average of normal controls. The pathological study of the cardiac biopsy sample showed no characteristic findings of Fabry disease. The immunohistochemistry for GL3 of the cardiac biopsy sample showed no positive cells. CONCLUSION: Although the E66Q mutation reduced enzyme activity, the characteristic pathological findings of Fabry disease and the abnormal accumulation of GL3 were not detected in cardiac tissues. The E66Q mutation of the GLA gene is thought to be a functional polymorphism based on enzymatic and pathological studies.

Novel automated screening of age-related macular degeneration.

PURPOSE: To determine the objective and quantitative hyperspectral parameters for distinguishing between age-related macular degeneration (AMD) and a normal macula. METHODS: Near-infrared hyperspectral images were taken of 71 eyes of 62 AMD patients with exudative AMD and 21 eyes of 12 control subjects without AMD. The spatial information included a 480 × 321-pixel image in a 50° field located at the ocular fundus and a 720-950-nm-per-pixel reflectance spectrum. Macular vectors were determined as the average spectrum for each macula, and reference vectors were used as average macular vectors for healthy volunteers. Variations in vector length and angle were calculated based on comparison with the reference vector. The AMD differentiation index was a parameter that minimized the plot overlap between AMD patients and controls. RESULTS: Statistically significant differences between the AMD patients and controls were noted. Receiver-operating characteristic curve analysis revealed an area under the curve of 0.888. The appropriate threshold values were attained for the proposed discrimination index, including 68 % sensitivity, 95 % specificity and 74 % accuracy. CONCLUSIONS: This study presents a simplified diagnostic index for the determination of age-related macular degeneration based on near-infrared spectra.

A possible melanoma discrimination index based on hyperspectral data: a pilot study.

BACKGROUND: Early detection and proper excision of the primary lesions of malignant melanoma (MM) are crucial for reducing melanoma-related deaths. To support the early detection of melanoma, automated melanoma screening systems have been extensively studied and developed. In this article, we present a hyperspectral melanoma screening system and propose a possible melanoma discrimination index derived from the characteristics of the pigment molecules in the skin, both of which have been derived from hyperspectral data (HSD). METHODS: The index expresses the disordered nature of each lesion including variegation in color based on variation in spectral information obtained from each lesion. Performance of the index in discriminating melanomas from other pigmented skin lesions has been studied in five cases of melanoma (41 HSD sets), one case of Spitz nevus (13 HSD sets), 10 cases of seborrheic keratosis (78 HSD sets), three cases of basal cell carcinoma (16 HSD sets), and nine cases of melanocytic nevus (21 HSD sets), obtained from patients and volunteers, all of whom were Japanese. RESULTS: Performance of the index, which reflects the disordered nature of a lesion, discriminates melanomas with a sensitivity of 90%, a specificity of 84%, and an area under the receiver operating characteristic curve of 0.93, on resubstitution. CONCLUSION: An objective melanoma discrimination index at a molecular pigmentary level, derived from HSD, has been proposed, and its performance evaluated. This index was highly successful in discriminating MM from non-melanoma, although the statistical population was small.

Development of blood vessel viewer using multi spectral imaging technology.

Using multi spectral imaging technology, which enable us to simultaneously obtain spatial and wavelength information, we have newly developed blood vessel viewer. Our device is composed of the light source, the lens or the endoscope, the optical device, CCD camera, and PC. As an evaluation of our device, we measured the skin diffuse reflectance spectrum of the human hand. The blood vessel observed by our device was corresponding to an anatomy finding. However, even if Principal Component Analysis was used, the reflection of horny was not able to be removed completely.

Inspection of skin hemodynamics with hyperspectral camera.

Using hyperspectral imaging techniques, which enable us to simultaneously obtain spatial and wavelength information, we have improved upon the newly developed hyperspectral camera to develop a method of observing changes in skin melanin levels and hemodynamics over time. As an evaluation of this method, we measured the skin diffuse reflectance spectrum of the human middle finger in an experiment of blood flow blockage in the brachial region. The changes in skin hemodynamics observed through this method match the behavior expected based on clinical knowledge, and also show an extremely high correlation with results obtained by the Erythema Index, which is used to make similar calculations from a limited number of wavelengths.

Noninvasive measurement of total hemoglobin and hemoglobin derivatives using multiwavelength pulse spectrophotometry -In vitro study with a mock circulatory system.

Total hemoglobin (tHb), carboxyhemoglobin (COHb), and methemoglobin (MetHb) are usually measured with a CO-oximeter. Noninvasive and continuous measurement of these blood components is expected to decrease the pain of a patient. Therefore, we developed an instrument to measure oxygen saturation (SpO(2), tHb, COHb, and MetHb non invasively. Multiwavelength LED (600, 625, 660, 760, 800, 940, and 1300 nm) and a combined detector (Si, InGaAs) were built into the instrument (Seven wavelengths transparent pulse spectrophotometer). We used the Waseda mock circulatory system, which can simulate blood circulation in tissues and generate a pulse wave mechanically, to estimate the instrument's performance. Furthermore we proposed new calculation formula including DC components of optical density (this method). Under conditions without any change of other components, the mean error +/- standard deviation between this method and the CO-oximetry were SaO(2)=0.0 +/- 1.4%, tHb=0.0 +/- 0.0 g/dl, COHb=0.0 +/- 2.0%, and MetHb=0.0 +/- 0.3%. When the concentration of other components was changed, this method showed mean errors and standard deviations of SaO(2)=0.2 +/- 1.6%, tHb=0.0 +/- 0.4 g/dl, COHb=0.5 +/- 4.1%, and MetHb=0.0 +/- 0.3%