Pub Date : 2022-06-22DOI: 10.1109/MeMeA54994.2022.9856529
J. Panić, V. Giannini, Arianna Defeudis, D. Regge, G. Balestra, S. Rosati
The use of Deep Learning (DL) algorithms in the medical imaging field is increasing in recent years. However, they require the selection of a set of parameters to properly perform. In this study we evaluated the impact of three factors (the construction of the training set, the number of network layers and the loss function) on the performance of a U-Net system in the segmentation of Locally Advanced Rectal Cancer (LARC) on Magnetic Resonance Imaging (MRI). Images from 3 different institutions and 4 different scanners were used to this scope, for a total of 100 patients. All images underwent a pre-processing step to normalize and to highlight the tumoral area. The sequences of two scanners were used to construct the networks while the remaining sequences were employed for validating the best performing systems. From our results, it emerged that Dice Similarity Coefficient is not affected by any of the evaluated factors. Conversely, the choice of loss function could bias the results towards either precision or recall and, thus, it should be properly performed according to the scope of the network. Moreover, a slightly improvement of the performances was observed using a training set based on clustering, maybe due to a better representation of the heterogeneity characterizing medical images.
{"title":"Impact of network parameters on a U-Net based system for rectal cancer segmentation on MR images","authors":"J. Panić, V. Giannini, Arianna Defeudis, D. Regge, G. Balestra, S. Rosati","doi":"10.1109/MeMeA54994.2022.9856529","DOIUrl":"https://doi.org/10.1109/MeMeA54994.2022.9856529","url":null,"abstract":"The use of Deep Learning (DL) algorithms in the medical imaging field is increasing in recent years. However, they require the selection of a set of parameters to properly perform. In this study we evaluated the impact of three factors (the construction of the training set, the number of network layers and the loss function) on the performance of a U-Net system in the segmentation of Locally Advanced Rectal Cancer (LARC) on Magnetic Resonance Imaging (MRI). Images from 3 different institutions and 4 different scanners were used to this scope, for a total of 100 patients. All images underwent a pre-processing step to normalize and to highlight the tumoral area. The sequences of two scanners were used to construct the networks while the remaining sequences were employed for validating the best performing systems. From our results, it emerged that Dice Similarity Coefficient is not affected by any of the evaluated factors. Conversely, the choice of loss function could bias the results towards either precision or recall and, thus, it should be properly performed according to the scope of the network. Moreover, a slightly improvement of the performances was observed using a training set based on clustering, maybe due to a better representation of the heterogeneity characterizing medical images.","PeriodicalId":106228,"journal":{"name":"2022 IEEE International Symposium on Medical Measurements and Applications (MeMeA)","volume":"43 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124412855","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-06-22DOI: 10.1109/MeMeA54994.2022.9856461
D. Borzelli, A. d’Avella, S. Gurgone, L. Gastaldi
EMG-driven robotic devices require the estimation of the forces exerted by the human operator from muscle activity. Approximating the relation between EMG and force with a linear mapping may be accurate enough for numerous real-time applications, such as controlling exoskeletons or prostheses. However, while a linear mapping from the EMG activity to endpoint force may be identified by minimizing the error without any constraint, introducing some constraints may be helpful to determine a mapping which is more anatomically accurate. The presence of noise and the muscle redundancy may introduce errors in the estimation achieved by the unconstrained optimization. Contrarily, anatomical constraints, estimated from an accurate musculoskeletal model, would limit the effect of noise, but they would increase the algorithm complexity and its computational costs. This study compares the two algorithms (unconstrained and constrained) for the estimation of the forces exerted by a human participant from the EMG activity of several upper limb muscles. The two algorithms were tested on data collected during an isometric force generation task performed during multiple sessions spanning two days. Accuracy and consistency across sessions of the reconstructed forces were assessed. Data showed that the unconstrained algorithm allowed for a better reconstruction of the exerted forces, but the constrained mapping is more robust across sessions. Further studies will investigate which of the two algorithms reconstruct a mapping perceived by the participants as more natural during EMG-driven control.
{"title":"Unconstrained and constrained estimation of a linear EMG-to-force mapping during isometric force generation","authors":"D. Borzelli, A. d’Avella, S. Gurgone, L. Gastaldi","doi":"10.1109/MeMeA54994.2022.9856461","DOIUrl":"https://doi.org/10.1109/MeMeA54994.2022.9856461","url":null,"abstract":"EMG-driven robotic devices require the estimation of the forces exerted by the human operator from muscle activity. Approximating the relation between EMG and force with a linear mapping may be accurate enough for numerous real-time applications, such as controlling exoskeletons or prostheses. However, while a linear mapping from the EMG activity to endpoint force may be identified by minimizing the error without any constraint, introducing some constraints may be helpful to determine a mapping which is more anatomically accurate. The presence of noise and the muscle redundancy may introduce errors in the estimation achieved by the unconstrained optimization. Contrarily, anatomical constraints, estimated from an accurate musculoskeletal model, would limit the effect of noise, but they would increase the algorithm complexity and its computational costs. This study compares the two algorithms (unconstrained and constrained) for the estimation of the forces exerted by a human participant from the EMG activity of several upper limb muscles. The two algorithms were tested on data collected during an isometric force generation task performed during multiple sessions spanning two days. Accuracy and consistency across sessions of the reconstructed forces were assessed. Data showed that the unconstrained algorithm allowed for a better reconstruction of the exerted forces, but the constrained mapping is more robust across sessions. Further studies will investigate which of the two algorithms reconstruct a mapping perceived by the participants as more natural during EMG-driven control.","PeriodicalId":106228,"journal":{"name":"2022 IEEE International Symposium on Medical Measurements and Applications (MeMeA)","volume":"37 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127847278","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-06-22DOI: 10.1109/MeMeA54994.2022.9856424
R. Morello, A. Sagaidachnyi, D. Quattrone
Pain is a debilitating condition affecting about 20% of adults in the world. It can be considered as a warning mechanism or the response of human body to alert about a harmful state. It involves complex neuronal processes and it is considered as a personal experience with a relevant subjective component. In specific conditions, pain can be so debilitating that it alters feelings and attitudes. So pain has important physical, psychological and social consequences and it can affect the quality of life. In absence of suitable and prompt treatments, the immune system can be compromised and pain sensation can interfere with the person ability to eat, concentrate, sleep, or interact with others. Consequently, the prompt and accurate pain assessment is essential for expediting therapeutic administration. Today, assessment, management and treatment of chronic pain are still challenging goals for researchers and clinicians. Algologists operate in absence of standard objective detection tools for pain assessment. As it remains confined to a subjective experience, pain has, like gold standard for its assessment, the patients' self-report. So it is clear the need to define new objective assessing tools. In this paper, the authors propose the use of the active thermography to analyse the neurogenic inflammatory response which characterizes nociceptive pain. Preliminary results of this feasibility study are here reported. Results have shown the potentiality of thermography to be a screening biomarker of the mechanism responsible of the abnormalities in sympathetic nervous system due to pain. In fact, a clear abnormality in the thermal response of subjects suffering from pain has been recorded in several cases. Asymmetries in temperature distribution of the two limbs have been observed, and in specific cases, an unbalanced trend of the thermoregulatory response to external thermal stimuli has been even highlighted. Although, these temperature differences have not been observed in all subjects with the same intensity and frequency, data provide evidence on the potentiality about the use of thermography to analyse pain mechanism.
{"title":"Feasibility Study of Pain Assessment by using Thermography","authors":"R. Morello, A. Sagaidachnyi, D. Quattrone","doi":"10.1109/MeMeA54994.2022.9856424","DOIUrl":"https://doi.org/10.1109/MeMeA54994.2022.9856424","url":null,"abstract":"Pain is a debilitating condition affecting about 20% of adults in the world. It can be considered as a warning mechanism or the response of human body to alert about a harmful state. It involves complex neuronal processes and it is considered as a personal experience with a relevant subjective component. In specific conditions, pain can be so debilitating that it alters feelings and attitudes. So pain has important physical, psychological and social consequences and it can affect the quality of life. In absence of suitable and prompt treatments, the immune system can be compromised and pain sensation can interfere with the person ability to eat, concentrate, sleep, or interact with others. Consequently, the prompt and accurate pain assessment is essential for expediting therapeutic administration. Today, assessment, management and treatment of chronic pain are still challenging goals for researchers and clinicians. Algologists operate in absence of standard objective detection tools for pain assessment. As it remains confined to a subjective experience, pain has, like gold standard for its assessment, the patients' self-report. So it is clear the need to define new objective assessing tools. In this paper, the authors propose the use of the active thermography to analyse the neurogenic inflammatory response which characterizes nociceptive pain. Preliminary results of this feasibility study are here reported. Results have shown the potentiality of thermography to be a screening biomarker of the mechanism responsible of the abnormalities in sympathetic nervous system due to pain. In fact, a clear abnormality in the thermal response of subjects suffering from pain has been recorded in several cases. Asymmetries in temperature distribution of the two limbs have been observed, and in specific cases, an unbalanced trend of the thermoregulatory response to external thermal stimuli has been even highlighted. Although, these temperature differences have not been observed in all subjects with the same intensity and frequency, data provide evidence on the potentiality about the use of thermography to analyse pain mechanism.","PeriodicalId":106228,"journal":{"name":"2022 IEEE International Symposium on Medical Measurements and Applications (MeMeA)","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132333043","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-06-22DOI: 10.1109/MeMeA54994.2022.9856472
A. Anusha, A. G. Ramakrishnan, A. Adarsh, Kanishka Sharma, G. P. Kumar
Recent years have seen a wealth of literature increasingly recognizing the concept that breathing rhythms entrain the activity of the human brain. However, the mechanisms underlying this phenomenon, and the extent to which rhythmic brain activity is modulated by breathing are not fully understood at the moment. The study reported herein is a preliminary step towards that goal. The variations in the electroencephalogram (EEG) based functional connectivity (FC) of the human brain during normal breathing, and voluntary breath-hold has been investigated and reported here. An experimental protocol involving breathing and breath-hold sessions, synchronized to a visual-metronome was designed and administered on 20 healthy subjects (9 females and 11 males within a range of 23–60 years). EEG data were collected from all subjects during breathing and breath-hold sessions using the 64 channel eego™mylab system from ANT Neuro. Further, FC was estimated on brain hemispheres and 7 cortical regions for 5 specific EEG bands, and variations were examined statistically. The observations illustrated that the brain FC exhibits a hemispherical symmetry during breath-hold in the delta and alpha bands. Synchronization of neuronal assemblies in different cortical regions of the brain was found to be higher in low-frequency EEG bands and lower in high-frequency EEG bands. Furthermore, the study also revealed that gamma-band FC of the pre-frontal cortex could distinctly identify an inhale-hold from exhale-hold.
{"title":"Effects of Breathing and Breath-hold on Brain Functional Connectivity","authors":"A. Anusha, A. G. Ramakrishnan, A. Adarsh, Kanishka Sharma, G. P. Kumar","doi":"10.1109/MeMeA54994.2022.9856472","DOIUrl":"https://doi.org/10.1109/MeMeA54994.2022.9856472","url":null,"abstract":"Recent years have seen a wealth of literature increasingly recognizing the concept that breathing rhythms entrain the activity of the human brain. However, the mechanisms underlying this phenomenon, and the extent to which rhythmic brain activity is modulated by breathing are not fully understood at the moment. The study reported herein is a preliminary step towards that goal. The variations in the electroencephalogram (EEG) based functional connectivity (FC) of the human brain during normal breathing, and voluntary breath-hold has been investigated and reported here. An experimental protocol involving breathing and breath-hold sessions, synchronized to a visual-metronome was designed and administered on 20 healthy subjects (9 females and 11 males within a range of 23–60 years). EEG data were collected from all subjects during breathing and breath-hold sessions using the 64 channel eego™mylab system from ANT Neuro. Further, FC was estimated on brain hemispheres and 7 cortical regions for 5 specific EEG bands, and variations were examined statistically. The observations illustrated that the brain FC exhibits a hemispherical symmetry during breath-hold in the delta and alpha bands. Synchronization of neuronal assemblies in different cortical regions of the brain was found to be higher in low-frequency EEG bands and lower in high-frequency EEG bands. Furthermore, the study also revealed that gamma-band FC of the pre-frontal cortex could distinctly identify an inhale-hold from exhale-hold.","PeriodicalId":106228,"journal":{"name":"2022 IEEE International Symposium on Medical Measurements and Applications (MeMeA)","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133724458","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-06-22DOI: 10.1109/MeMeA54994.2022.9856412
S. Arlati, N. Keijsers, G. Paolini, G. Ferrigno, M. Sacco
Immersive virtual reality (VR) represents a viable technology to support rehabilitation and promote the recovery of upper limb functions after stroke. Nonetheless, it has not been determined yet if VR can elicit movements that share the same kinematic characteristics of those occurring in the real world (RW), thus positively impacting arm use in daily life. A previous study enrolling young adults showed promising results: joints' ranges of motion were preserved, although movement times were longer and peak velocity lower in VR. Starting from these results, this work aimed at comparing young and older adults' (i.e., an age-matched sample to stroke survivors) upper limb kinematics while performing aimed movements in RW and immersive VR. The presented study was a within-subject repeated-measures design in which participants had to reach, grasp and transport grocery items from a simplified supermarket shelf unit. The VR condition was performed using an HTC Vive head-mounted display; its controller was used to interact with virtual objects. Three conditions were tested: VR, RW, and RW while holding the controller (RWC, to account for carrying a weight). Ten healthy young adults $(26.7pm 5.46$ and three older adults $(69.0pm 2.0)$ were enrolled. The collected data showed that older adults moved slower, more curved, and reached lower peak velocity during both reaching and transfer in VR compared to young adults. Arm ranges of motion seemed to be preserved, whereas thorax movements were different. We hypothesized that these differences might be dependent on age-related vision and cognitive decline, lack of familiarity with VR technology, and lack of force feedback. Further studies are needed to address these issues and confirm or reject our hypotheses.
{"title":"Age-related differences in the kinematics of aimed movements in immersive virtual reality: a preliminary study","authors":"S. Arlati, N. Keijsers, G. Paolini, G. Ferrigno, M. Sacco","doi":"10.1109/MeMeA54994.2022.9856412","DOIUrl":"https://doi.org/10.1109/MeMeA54994.2022.9856412","url":null,"abstract":"Immersive virtual reality (VR) represents a viable technology to support rehabilitation and promote the recovery of upper limb functions after stroke. Nonetheless, it has not been determined yet if VR can elicit movements that share the same kinematic characteristics of those occurring in the real world (RW), thus positively impacting arm use in daily life. A previous study enrolling young adults showed promising results: joints' ranges of motion were preserved, although movement times were longer and peak velocity lower in VR. Starting from these results, this work aimed at comparing young and older adults' (i.e., an age-matched sample to stroke survivors) upper limb kinematics while performing aimed movements in RW and immersive VR. The presented study was a within-subject repeated-measures design in which participants had to reach, grasp and transport grocery items from a simplified supermarket shelf unit. The VR condition was performed using an HTC Vive head-mounted display; its controller was used to interact with virtual objects. Three conditions were tested: VR, RW, and RW while holding the controller (RWC, to account for carrying a weight). Ten healthy young adults $(26.7pm 5.46$ and three older adults $(69.0pm 2.0)$ were enrolled. The collected data showed that older adults moved slower, more curved, and reached lower peak velocity during both reaching and transfer in VR compared to young adults. Arm ranges of motion seemed to be preserved, whereas thorax movements were different. We hypothesized that these differences might be dependent on age-related vision and cognitive decline, lack of familiarity with VR technology, and lack of force feedback. Further studies are needed to address these issues and confirm or reject our hypotheses.","PeriodicalId":106228,"journal":{"name":"2022 IEEE International Symposium on Medical Measurements and Applications (MeMeA)","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128100930","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-06-22DOI: 10.1109/MeMeA54994.2022.9856457
Daniel G. Kyrollos, K. Greenwood, J. Harrold, J. Green
This paper explores the use of two types of transfer learning for the task of neonatal head localization from pressure images: 1) The pretrained CNN portion of the PressureNet model, a deep learning model that estimates adult pose given a pressure image, is used for transfer learning for a neonatal population. 2) Annotation of the training patient head locations was completed in the RGB image domain, then transferred to the pressure image domain of application. A multi-modal neonatal patient dataset suitable for this task was used. Data was simultaneously collected from a RGB-D video camera placed above the patient and a pressure sensitive mat (PSM) beneath the patient. Geometric transforms were used to achieve spatial registration between the video image plane and the PSM plane. Patient localization is important in the application of noncontact monitoring for vital sign estimation and movement detection. In testing on unseen patients, 54% of detections made by the object detection model achieved an IoU of 0.5 or greater. This is higher than the accuracy (33%) achieved using a pre-trained ResNet model trained with pressure images converted to RGB. This study demonstrates the potential for cross-domain transfer learning between RGB image and PSM domains.
{"title":"Transfer Learning Approaches for Neonate Head Localization from Pressure Images","authors":"Daniel G. Kyrollos, K. Greenwood, J. Harrold, J. Green","doi":"10.1109/MeMeA54994.2022.9856457","DOIUrl":"https://doi.org/10.1109/MeMeA54994.2022.9856457","url":null,"abstract":"This paper explores the use of two types of transfer learning for the task of neonatal head localization from pressure images: 1) The pretrained CNN portion of the PressureNet model, a deep learning model that estimates adult pose given a pressure image, is used for transfer learning for a neonatal population. 2) Annotation of the training patient head locations was completed in the RGB image domain, then transferred to the pressure image domain of application. A multi-modal neonatal patient dataset suitable for this task was used. Data was simultaneously collected from a RGB-D video camera placed above the patient and a pressure sensitive mat (PSM) beneath the patient. Geometric transforms were used to achieve spatial registration between the video image plane and the PSM plane. Patient localization is important in the application of noncontact monitoring for vital sign estimation and movement detection. In testing on unseen patients, 54% of detections made by the object detection model achieved an IoU of 0.5 or greater. This is higher than the accuracy (33%) achieved using a pre-trained ResNet model trained with pressure images converted to RGB. This study demonstrates the potential for cross-domain transfer learning between RGB image and PSM domains.","PeriodicalId":106228,"journal":{"name":"2022 IEEE International Symposium on Medical Measurements and Applications (MeMeA)","volume":"9 1-2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131723221","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-06-22DOI: 10.1109/MeMeA54994.2022.9856491
Giorgia Fiori, A. Scorza, M. Schmid, J. Galo, S. Conforto, S. Sciuto
Quality Assessment (QA) of ultrasound (US) equipment is of primary importance since US diagnostic systems are used in a wide range of medical applications. Among the recommended test parameters, maximum depth of penetration, local dynamic range and spatial resolution are usually estimated in the literature through the Gray Scale Mapping Function (GSMF) that, for some methods, requires the US system's gain to be provided in dB. Since many US systems in the market provide the gain in arbitrary units (au), a novel automatic method for the assessment of the gain conversion factor to dB has been proposed and investigated in the present study. According to the definition, if the diagnostic system provides the overall gain in au, the abovementioned factor is the conversion unit from au to dB, while it is a dimensionless coefficient if the gain is directly given in dB. Data have been collected on a gray scale US phantom displaying the contrast targets at three different depths as well as by varying both the operating frequency of the phased array probe used and the dynamic range settings. Based on the promising preliminary results, further studies will be carried out on a higher number of diagnostic systems and probe models to improve the automatic method and deepen the method uncertainty investigation.
{"title":"A novel method for the gain conversion factor estimation in quality assessment of ultrasound diagnostic systems","authors":"Giorgia Fiori, A. Scorza, M. Schmid, J. Galo, S. Conforto, S. Sciuto","doi":"10.1109/MeMeA54994.2022.9856491","DOIUrl":"https://doi.org/10.1109/MeMeA54994.2022.9856491","url":null,"abstract":"Quality Assessment (QA) of ultrasound (US) equipment is of primary importance since US diagnostic systems are used in a wide range of medical applications. Among the recommended test parameters, maximum depth of penetration, local dynamic range and spatial resolution are usually estimated in the literature through the Gray Scale Mapping Function (GSMF) that, for some methods, requires the US system's gain to be provided in dB. Since many US systems in the market provide the gain in arbitrary units (au), a novel automatic method for the assessment of the gain conversion factor to dB has been proposed and investigated in the present study. According to the definition, if the diagnostic system provides the overall gain in au, the abovementioned factor is the conversion unit from au to dB, while it is a dimensionless coefficient if the gain is directly given in dB. Data have been collected on a gray scale US phantom displaying the contrast targets at three different depths as well as by varying both the operating frequency of the phased array probe used and the dynamic range settings. Based on the promising preliminary results, further studies will be carried out on a higher number of diagnostic systems and probe models to improve the automatic method and deepen the method uncertainty investigation.","PeriodicalId":106228,"journal":{"name":"2022 IEEE International Symposium on Medical Measurements and Applications (MeMeA)","volume":"67 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131875958","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-06-22DOI: 10.1109/MeMeA54994.2022.9856589
Arianna Defeudis, J. Panić, Walter Guzzinati, L. Pusceddu, L. Vassallo, D. Regge, V. Giannini
The aim of this study is to present a fully automatic deep learning algorithm to segment liver Colorectal cancer metastases (lmCRC) on CT images, based on a U-Net structure, comparing nets with and without the transfer learning approach. This is a bi-centric study, enrolling patients who underwent CT exam before (baseline) and after first-line therapy (TP1). Patients were divided into training (using a portion of baseline sequences from both centers) to train the DL model, and two validation sets: one with baseline (valB), and one with TP1 (valTP1) sequences. The reference standard for the automatic segmentations was defined by the manual segmentations performed by an experienced radiologist on the portal phase of the baseline and TP1 CT exam. The best performing model obtained Dice Similarity Coefficient (DSC) of $0.68pm 0.24$, Precision (Pr) of $0.74pm 0.27$, Recall (Re) of $0.73pm 0.26$, Detection Rate (DR) of 93% on the valB, and DSC of $0.61pm 0.28$, Pr of $0.68pm 0.31$, Re of $0.65pm 0.29$ and DR of 88% on the valTP1. These encouraging results, if confirmed on larger dataset, might provide a reliable and robust tool that can be used as first step of future radiomics analyses aimed at predicting response to therapy, improving the management of lmCRC patients.
{"title":"A Deep Learning model to segment liver metastases on CT images acquired at different time-points during chemotherapy","authors":"Arianna Defeudis, J. Panić, Walter Guzzinati, L. Pusceddu, L. Vassallo, D. Regge, V. Giannini","doi":"10.1109/MeMeA54994.2022.9856589","DOIUrl":"https://doi.org/10.1109/MeMeA54994.2022.9856589","url":null,"abstract":"The aim of this study is to present a fully automatic deep learning algorithm to segment liver Colorectal cancer metastases (lmCRC) on CT images, based on a U-Net structure, comparing nets with and without the transfer learning approach. This is a bi-centric study, enrolling patients who underwent CT exam before (baseline) and after first-line therapy (TP1). Patients were divided into training (using a portion of baseline sequences from both centers) to train the DL model, and two validation sets: one with baseline (valB), and one with TP1 (valTP1) sequences. The reference standard for the automatic segmentations was defined by the manual segmentations performed by an experienced radiologist on the portal phase of the baseline and TP1 CT exam. The best performing model obtained Dice Similarity Coefficient (DSC) of $0.68pm 0.24$, Precision (Pr) of $0.74pm 0.27$, Recall (Re) of $0.73pm 0.26$, Detection Rate (DR) of 93% on the valB, and DSC of $0.61pm 0.28$, Pr of $0.68pm 0.31$, Re of $0.65pm 0.29$ and DR of 88% on the valTP1. These encouraging results, if confirmed on larger dataset, might provide a reliable and robust tool that can be used as first step of future radiomics analyses aimed at predicting response to therapy, improving the management of lmCRC patients.","PeriodicalId":106228,"journal":{"name":"2022 IEEE International Symposium on Medical Measurements and Applications (MeMeA)","volume":"252 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134236674","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-06-22DOI: 10.1109/MeMeA54994.2022.9856528
L. E. Sebar, E. Angelini, A. Baldi, A. Comba, M. Parvis, S. Grassini
The employment of innovative all-ceramic materi-als and adhesive cement, as well as the development of new bonding procedures, allow clinicians to use minimally invasive approaches in conservative restorations. In particular, dual-cure cement allows for obtaining higher aesthetic and functional results. However, the reduced light transmission through ceramic materials could prevent the proper curing and affect the adhesion of these materials to the tooth surface. In this context, the development of an accurate measurement methodology to assess the extent of polymerization of dental resin-based luting cement and to correlate the conversion degree with the mechanical properties is of particular importance from the clinical and scientific point of view. A measurement approach that exploits Raman Spectroscopy and nano-hardness measurements is hereby proposed. In particular, in this study, two different light-curing protocols are employed on a dual-cure luting cement, usually used for the full-crown restoration of single-rooted teeth. The effect of different times and tack-curing steps on the polymerization shrinkage of resin-based luting cement is investigated. The pre-liminary results allow concluding that both curing protocols lead to a good polymerization, without significant differences in the degree of conversion along the cement-tooth interfacial surface, as proved by the almost constant ratio of the Raman vibration characteristic peaks. However, the nanoindentation modulus was lower in the case of the tack-cured protocol.
{"title":"Nanoindentation and Raman spectroscopy measurements on dual-cure luting cement for dental conservative restoration","authors":"L. E. Sebar, E. Angelini, A. Baldi, A. Comba, M. Parvis, S. Grassini","doi":"10.1109/MeMeA54994.2022.9856528","DOIUrl":"https://doi.org/10.1109/MeMeA54994.2022.9856528","url":null,"abstract":"The employment of innovative all-ceramic materi-als and adhesive cement, as well as the development of new bonding procedures, allow clinicians to use minimally invasive approaches in conservative restorations. In particular, dual-cure cement allows for obtaining higher aesthetic and functional results. However, the reduced light transmission through ceramic materials could prevent the proper curing and affect the adhesion of these materials to the tooth surface. In this context, the development of an accurate measurement methodology to assess the extent of polymerization of dental resin-based luting cement and to correlate the conversion degree with the mechanical properties is of particular importance from the clinical and scientific point of view. A measurement approach that exploits Raman Spectroscopy and nano-hardness measurements is hereby proposed. In particular, in this study, two different light-curing protocols are employed on a dual-cure luting cement, usually used for the full-crown restoration of single-rooted teeth. The effect of different times and tack-curing steps on the polymerization shrinkage of resin-based luting cement is investigated. The pre-liminary results allow concluding that both curing protocols lead to a good polymerization, without significant differences in the degree of conversion along the cement-tooth interfacial surface, as proved by the almost constant ratio of the Raman vibration characteristic peaks. However, the nanoindentation modulus was lower in the case of the tack-cured protocol.","PeriodicalId":106228,"journal":{"name":"2022 IEEE International Symposium on Medical Measurements and Applications (MeMeA)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129001575","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-06-22DOI: 10.1109/MeMeA54994.2022.9856442
S. Grassini, L. E. Sebar, A. Baldi, A. Comba, E. Angelini, E. Berutti
The paper deals with a measuring approach based on Raman Spectroscopy and micro-CT imaging for correlating the degree of conversion of bulk-fill composites to the contraction shrinkage and consequently to the internal gap formation in high c-factor dental cavities. The developed study was performed on extracted molars in which a first-class cavity was prepared. A micro-CT scan was performed before and after composite lightcuring to tridimensionally measure the interfacial gap between the composite material and the cavity walls. After the complete polymerization of the composite, each sample was sectioned vertically to expose the lateral surface of the restorative material. Raman Spectroscopy measurements were performed along the cross-section of the cavity filled with the restorative material, every 0.5 mm from the occlusal surface. The obtained results showed a minimal gap opening after light-curing and a degree of conversion which was not affected by the bulk-fill composite thickness. Thanks to the 3D rendering, it should be observed that gaps were mostly concentrated at the cavity floor and despite the reduction in the degree of conversion detected in the deeper portions of the restoration, a three-dimensional opening of an interfacial gap was not observed. Therefore, it is possible to assume the presence of a correlation between the degree of conversion and the volumetric interfacial gap could. Further studies are actually in progress to compare these preliminary results with those obtained on other dental composite materials.
{"title":"Measurements for restorative dentistry: shrinkage and conversion degree of bulk-fill composites","authors":"S. Grassini, L. E. Sebar, A. Baldi, A. Comba, E. Angelini, E. Berutti","doi":"10.1109/MeMeA54994.2022.9856442","DOIUrl":"https://doi.org/10.1109/MeMeA54994.2022.9856442","url":null,"abstract":"The paper deals with a measuring approach based on Raman Spectroscopy and micro-CT imaging for correlating the degree of conversion of bulk-fill composites to the contraction shrinkage and consequently to the internal gap formation in high c-factor dental cavities. The developed study was performed on extracted molars in which a first-class cavity was prepared. A micro-CT scan was performed before and after composite lightcuring to tridimensionally measure the interfacial gap between the composite material and the cavity walls. After the complete polymerization of the composite, each sample was sectioned vertically to expose the lateral surface of the restorative material. Raman Spectroscopy measurements were performed along the cross-section of the cavity filled with the restorative material, every 0.5 mm from the occlusal surface. The obtained results showed a minimal gap opening after light-curing and a degree of conversion which was not affected by the bulk-fill composite thickness. Thanks to the 3D rendering, it should be observed that gaps were mostly concentrated at the cavity floor and despite the reduction in the degree of conversion detected in the deeper portions of the restoration, a three-dimensional opening of an interfacial gap was not observed. Therefore, it is possible to assume the presence of a correlation between the degree of conversion and the volumetric interfacial gap could. Further studies are actually in progress to compare these preliminary results with those obtained on other dental composite materials.","PeriodicalId":106228,"journal":{"name":"2022 IEEE International Symposium on Medical Measurements and Applications (MeMeA)","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123930559","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：