Browsing by Author "Jaskari, Joel"
Now showing 1 - 20 of 24
Results Per Page
Sort Options
Item Application of simultaneous uncertainty quantification and segmentation for oropharyngeal cancer use-case with Bayesian deep learning(Nature Portfolio, 2024-12) Sahlsten, Jaakko; Jaskari, Joel; Wahid, Kareem A.; Ahmed, Sara; Glerean, Enrico; He, Renjie; Kann, Benjamin H.; Mäkitie, Antti; Fuller, Clifton D.; Naser, Mohamed A.; Kaski, Kimmo; Department of Computer Science; Department of Neuroscience and Biomedical Engineering; Kaski Kimmo group; Department of Computer Science; University of Texas MD Anderson Cancer Center; Harvard Medical School; University of HelsinkiBackground Radiotherapy is a core treatment modality for oropharyngeal cancer (OPC), where the primary gross tumor volume (GTVp) is manually segmented with high interobserver variability. This calls for reliable and trustworthy automated tools in clinician workflow. Therefore, accurate uncertainty quantification and its downstream utilization is critical. Methods Here we propose uncertainty-aware deep learning for OPC GTVp segmentation, and illustrate the utility of uncertainty in multiple applications. We examine two Bayesian deep learning (BDL) models and eight uncertainty measures, and utilize a large multi-institute dataset of 292 PET/CT scans to systematically analyze our approach. Results We show that our uncertainty-based approach accurately predicts the quality of the deep learning segmentation in 86.6% of cases, identifies low performance cases for semi-automated correction, and visualizes regions of the scans where the segmentations likely fail. Conclusions Our BDL-based analysis provides a first-step towards more widespread implementation of uncertainty quantification in OPC GTVp segmentation.Item Artificial Intelligence for Radiation Oncology Applications Using Public Datasets(W.B. Saunders Ltd, 2022-10) Wahid, Kareem A.; Glerean, Enrico; Sahlsten, Jaakko; Jaskari, Joel; Kaski, Kimmo; Naser, Mohamed A.; He, Renjie; Mohamed, Abdallah S.R.; Fuller, Clifton D.; Department of Computer Science; Department of Neuroscience and Biomedical Engineering; Kaski Kimmo group; University of Texas MD Anderson Cancer Center; Department of Computer ScienceArtificial intelligence (AI) has exceptional potential to positively impact the field of radiation oncology. However, large curated datasets - often involving imaging data and corresponding annotations - are required to develop radiation oncology AI models. Importantly, the recent establishment of Findable, Accessible, Interoperable, Reusable (FAIR) principles for scientific data management have enabled an increasing number of radiation oncology related datasets to be disseminated through data repositories, thereby acting as a rich source of data for AI model building. This manuscript reviews the current and future state of radiation oncology data dissemination, with a particular emphasis on published imaging datasets, AI data challenges, and associated infrastructure. Moreover, we provide historical context of FAIR data dissemination protocols, difficulties in the current distribution of radiation oncology data, and recommendations regarding data dissemination for eventual utilization in AI models. Through FAIR principles and standardized approaches to data dissemination, radiation oncology AI research has nothing to lose and everything to gain.Item Auto-detection and segmentation of involved lymph nodes in HPV-associated oropharyngeal cancer using a convolutional deep learning neural network(Elsevier Ireland Ltd, 2022-09) Taku, Nicolette; Wahid, Kareem A.; van Dijk, Lisanne V.; Sahlsten, Jaakko; Jaskari, Joel; Kaski, Kimmo; Fuller, Clifton D.; Naser, Mohamed A.; Department of Computer Science; Kaski Kimmo group; University of Texas MD Anderson Cancer Center; University of Groningen; Department of Computer SciencePurpose: Segmentation of involved lymph nodes on head and neck computed tomography (HN-CT) scans is necessary for the radiotherapy planning of early-stage human papilloma virus (HPV) associated oropharynx cancers (OPC). We aimed to train a deep learning convolutional neural network (DL-CNN) to segment involved lymph nodes on HN-CT scans. Methods: Ground-truth segmentation of involved nodes was performed on pre-surgical HN-CT scans for 90 patients who underwent levels II-IV neck dissection for node-positive HPV-OPC (training/validation [n = 70] and testing [n = 20]). A 5-fold cross validation approach was used to train 5 DL-CNN sub-models based on a residual U-net architecture. Validation and testing segmentation masks were compared to ground-truth masks using predetermined metrics. A lymph auto-detection model to discriminate between “node-positive” and “node-negative” HN-CT scans was developed by thresholding segmentation model outputs and evaluated using the area under the receiver operating characteristic curve (AUC). Results: In the DL-CNN validation phase, all sub-models yielded segmentation masks with median Dice ≥ 0.90 and median volume similarity score of ≥ 0.95. In the testing phase, the DL-CNN produced consensus segmentation masks with median Dice of 0.92 (IQR, 0.89–0.95), median volume similarity of 0.97 (IQR, 0.94–0.99), and median Hausdorff distance of 4.52 mm (IQR, 1.22–8.38). The detection model achieved an AUC of 0.98. Conclusion: The results from this single-institution study demonstrate the successful automation of lymph node segmentation for patients with HPV-OPC using a DL-CNN. Future studies, including validation with an external dataset, are necessary to clarify its role in the larger radiation oncology treatment planning workflow.Item Classification of bone defects using natural and synthetic X-ray images(2021-05-17) Roy Choudhury, Akash; Kärkkäinen, Leo; Jaskari, Joel; Perustieteiden korkeakoulu; Särkkä, SimoIn this thesis, we study methods to reduce the amount of data needed to create deep learning models that can detect defects in bones from X-ray images. Detecting defects in bones from X-ray images and properly annotating the images is the paramount step when it comes to corrective surgeries of bones. Annotations or labels, such as radial inclination and volar tilt are measurements that are necessary for many corrective surgeries. Generating these annotations is an arduous and manual task for medical professionals. By being able to automate the process of generating these annotations, it will be possible to reduce a significant amount of labor of these professionals. Modern deep learning models are heavily reliant upon availability of a large amount of properly labeled data for their training. In this thesis, we experimented to find methods to create appropriate synthetic data that can be combined with natural data to train deep learning models. We designed three deep learning models to generate two different forms of annotations. The first goal was to use cycle consistent generative adversarial networks to create proper synthetic images. Then we used the synthetic images to improve classifier models that can detect defects in bones. In the end, we expanded the cycle consistent generative adversarial network so that it can accommodate three input domains instead of two and called it multi-cycleGAN. We used multi-cycleGAN to segment bones from natural X-ray images. Our experiments concluded that by adding proper synthetic images with natural images, we can improve the performance of classifiers significantly and circumvent the persistent issue of unavailability of data. However, the multi-cycleGAN model did not generate a very accurate segmentation of bones. It was able to segment bones of forearm better than bones of wrist. It was able to understand the overall shape and positioning of the wrists in X-ray images but it did not produce proper segmentations of the individual fingers.Item Comparison of deep learning segmentation and multigrader-annotated mandibular canals of multicenter CBCT scans(Nature Publishing Group, 2022-12) Järnstedt, Jorma; Sahlsten, Jaakko; Jaskari, Joel; Kaski, Kimmo; Mehtonen, Helena; Lin, Ziyuan; Hietanen, Ari; Sundqvist, Osku; Varjonen, Vesa; Mattila, Vesa; Prapayasotok, Sangsom; Nalampang, Sakarat; Department of Computer Science; Kaski Kimmo group; Probabilistic Machine Learning; Professorship Kaski Samuel; Department of Computer Science; Planmeca Oy; Chiang Mai University; Tampere UniversityDeep learning approach has been demonstrated to automatically segment the bilateral mandibular canals from CBCT scans, yet systematic studies of its clinical and technical validation are scarce. To validate the mandibular canal localization accuracy of a deep learning system (DLS) we trained it with 982 CBCT scans and evaluated using 150 scans of five scanners from clinical workflow patients of European and Southeast Asian Institutes, annotated by four radiologists. The interobserver variability was compared to the variability between the DLS and the radiologists. In addition, the generalisation of DLS to CBCT scans from scanners not used in the training data was examined to evaluate its out-of-distribution performance. The DLS had a statistically significant difference (p < 0.001) with lower variability to the radiologists with 0.74 mm than the interobserver variability of 0.77 mm and generalised to new devices with 0.63 mm, 0.67 mm and 0.87 mm (p < 0.001). For the radiologists’ consensus segmentation, used as a gold standard, the DLS showed a symmetric mean curve distance of 0.39 mm, which was statistically significantly different (p < 0.001) compared to those of the individual radiologists with values of 0.62 mm, 0.55 mm, 0.47 mm, and 0.42 mm. These results show promise towards integration of DLS into clinical workflow to reduce time-consuming and labour-intensive manual tasks in implantology.Item Deep learning auto-segmentation of cervical skeletal muscle for sarcopenia analysis in patients with head and neck cancer(Frontiers Research Foundation, 2022-07-28) Naser, Mohamed A.; Wahid, Kareem A.; Grossberg, Aaron J.; Olson, Brennan; Jain, Rishab; El-Habashy, Dina; Dede, Cem; Salama, Vivian; Abobakr, Moamen; Mohamed, Abdallah S.R.; He, Renjie; Jaskari, Joel; Sahlsten, Jaakko; Kaski, Kimmo; Fuller, Clifton D.; Department of Computer Science; Kaski Kimmo group; University of Texas MD Anderson Cancer Center; Oregon Health and Science University; Menoufia University; Department of Computer ScienceBackground/Purpose: Sarcopenia is a prognostic factor in patients with head and neck cancer (HNC). Sarcopenia can be determined using the skeletal muscle index (SMI) calculated from cervical neck skeletal muscle (SM) segmentations. However, SM segmentation requires manual input, which is time-consuming and variable. Therefore, we developed a fully-automated approach to segment cervical vertebra SM. Materials/Methods: 390 HNC patients with contrast-enhanced CT scans were utilized (300-training, 90-testing). Ground-truth single-slice SM segmentations at the C3 vertebra were manually generated. A multi-stage deep learning pipeline was developed, where a 3D ResUNet auto-segmented the C3 section (33 mm window), the middle slice of the section was auto-selected, and a 2D ResUNet auto-segmented the auto-selected slice. Both the 3D and 2D approaches trained five sub-models (5-fold cross-validation) and combined sub-model predictions on the test set using majority vote ensembling. Model performance was primarily determined using the Dice similarity coefficient (DSC). Predicted SMI was calculated using the auto-segmented SM cross-sectional area. Finally, using established SMI cutoffs, we performed a Kaplan-Meier analysis to determine associations with overall survival. Results: Mean test set DSC of the 3D and 2D models were 0.96 and 0.95, respectively. Predicted SMI had high correlation to the ground-truth SMI in males and females (r>0.96). Predicted SMI stratified patients for overall survival in males (log-rank p = 0.01) but not females (log-rank p = 0.07), consistent with ground-truth SMI. Conclusion: We developed a high-performance, multi-stage, fully-automated approach to segment cervical vertebra SM. Our study is an essential step towards fully-automated sarcopenia-related decision-making in patients with HNC.Item Deep learning for 3D cephalometric landmarking with heterogeneous multi-center CBCT dataset(Public Library of Science, 2024-06) Sahlsten, Jaakko; Järnstedt, Jorma; Jaskari, Joel; Naukkarinen, Hanna; Mahasantipiya, Phattaranant; Charuakkra, Arnon; Vasankari, Krista; Hietanen, Ari; Sundqvist, Osku; Lehtinen, Antti; Kaski, Kimmo; Department of Computer Science; Kaski Kimmo group; Department of Computer Science; Planmeca Oy; Chiang Mai University; Tampere UniversityCephalometric analysis is critically important and common procedure prior to orthodontic treatment and orthognathic surgery. Recently, deep learning approaches have been proposed for automatic 3D cephalometric analysis based on landmarking from CBCT scans. However, these approaches have relied on uniform datasets from a single center or imaging device but without considering patient ethnicity. In addition, previous works have considered a limited number of clinically relevant cephalometric landmarks and the approaches were computationally infeasible, both impairing integration into clinical workflow. Here our aim is to analyze the clinical applicability of a light-weight deep learning neural network for fast localization of 46 clinically significant cephalometric landmarks with multi-center, multi-ethnic, and multi-device data consisting of 309 CBCT scans from Finnish and Thai patients. The localization performance of our approach resulted in the mean distance of 1.99 ± 1.55 mm for the Finnish cohort and 1.96 ± 1.25 mm for the Thai cohort. This performance turned out to be clinically significant i.e., ≤ 2 mm with 61.7% and 64.3% of the landmarks with Finnish and Thai cohorts, respectively. Furthermore, the estimated landmarks were used to measure cephalometric characteristics successfully i.e., with ≤ 2 mm or ≤ 2̊ error, on 85.9% of the Finnish and 74.4% of the Thai cases. Between the two patient cohorts, 33 of the landmarks and all cephalometric characteristics had no statistically significant difference (p < 0.05) measured by the Mann-Whitney U test with Benjamini–Hochberg correction. Moreover, our method is found to be computationally light, i.e., providing the predictions with the mean duration of 0.77 s and 2.27 s with single machine GPU and CPU computing, respectively. Our findings advocate for the inclusion of this method into clinical settings based on its technical feasibility and robustness across varied clinical datasets.Item Deep Learning Fundus Image Analysis for Diabetic Retinopathy and Macular Edema Grading(NATURE PUBLISHING GROUP, 2019-12-01) Sahlsten, Jaakko; Jaskari, Joel; Kivinen, Jyri; Turunen, Lauri; Jaanio, Esa; Hietala, Kustaa; Kaski, Kimmo; Department of Computer Science; Kaski Kimmo group; Centre of Excellence in Computational Inference, COIN; Department of Computer Science; Digifundus Ltd.; Central Finland Central HospitalDiabetes is a globally prevalent disease that can cause visible microvascular complications such as diabetic retinopathy and macular edema in the human eye retina, the images of which are today used for manual disease screening and diagnosis. This labor-intensive task could greatly benefit from automatic detection using deep learning technique. Here we present a deep learning system that identifies referable diabetic retinopathy comparably or better than presented in the previous studies, although we use only a small fraction of images (<1/4) in training but are aided with higher image resolutions. We also provide novel results for five different screening and clinical grading systems for diabetic retinopathy and macular edema classification, including state-of-the-art results for accurately classifying images according to clinical five-grade diabetic retinopathy and for the first time for the four-grade diabetic macular edema scales. These results suggest, that a deep learning system could increase the cost-effectiveness of screening and diagnosis, while attaining higher than recommended performance, and that the system could be applied in clinical examinations requiring finer grading.Item Deep Learning Method for Mandibular Canal Segmentation in Dental Cone Beam Computed Tomography Volumes(Nature Publishing Group, 2020-12-01) Jaskari, Joel; Sahlsten, Jaakko; Järnstedt, Jorma; Mehtonen, Helena; Karhu, Kalle; Sundqvist, Osku; Hietanen, Ari; Varjonen, Vesa; Mattila, Vesa; Kaski, Kimmo; Department of Computer Science; Kaski Kimmo group; Department of Computer Science; Planmeca Oy; Tampere University HospitalAccurate localisation of mandibular canals in lower jaws is important in dental implantology, in which the implant position and dimensions are currently determined manually from 3D CT images by medical experts to avoid damaging the mandibular nerve inside the canal. Here we present a deep learning system for automatic localisation of the mandibular canals by applying a fully convolutional neural network segmentation on clinically diverse dataset of 637 cone beam CT volumes, with mandibular canals being coarsely annotated by radiologists, and using a dataset of 15 volumes with accurate voxel-level mandibular canal annotations for model evaluation. We show that our deep learning model, trained on the coarsely annotated volumes, localises mandibular canals of the voxel-level annotated set, highly accurately with the mean curve distance and average symmetric surface distance being 0.56 mm and 0.45 mm, respectively. These unparalleled accurate results highlight that deep learning integrated into dental implantology workflow could significantly reduce manual labour in mandibular canal annotations.Item Deep Learning Transformers in Diabetic Retinopathy Detection(2023-12-11) Kivinen, Viivi; Kaski, Kimmo; Jaskari, Joel; Perustieteiden korkeakoulu; Marttinen, PekkaDiabetic retinopathy is a progressive ocular disease linked with diabetes mellitus, which can lead to blindness. The constantly high blood glucose levels caused by diabetes damage the vascular system, which can be seen e.g. as damaged retinal blood vessels. Diabetic retinopathy can be detected by inspecting fundus images for any visual changes. This time-consuming process, executed by ophthalmologists, can be automated by creating efficient computer vision algorithms. Convolutional Neural Networks (CNN) have been established as the state-of-the-art architecture in computer vision tasks. However, an architecture based on the self-attention mechanism, called Transformer, has been recently utilized in computer vision with promising results, albeit requiring significantly larger training datasets and computational resources than traditional CNNs. In this thesis, three Transformer architectures, i.e., Swin, Vision Transformer-B/16 and Vision Transformer-B/8 are compared to a state-of-the-art CNN model called EfficientNet-B6, in the detection of the severity of diabetic retinopathy from patients’ fundus images. The results of this thesis indicate that the Transformer architecture has potential to improve upon the CNN in diabetic retinopathy detection. However, the amount of training data and required computational resources are challenges yet to be solved.Item Development and Validation of an Automated Image-Based Deep Learning Platform for Sarcopenia Assessment in Head and Neck Cancer(American Medical Association, 2023-08-01) Ye, Zezhong; Saraf, Anurag; Ravipati, Yashwanth; Hoebers, Frank; Catalano, Paul J.; Zha, Yining; Zapaishchykova, Anna; Likitlersuang, Jirapat; Guthier, Christian; Tishler, Roy B.; Schoenfeld, Jonathan D.; Margalit, Danielle N.; Haddad, Robert I.; Mak, Raymond H.; Naser, Mohamed; Wahid, Kareem A.; Sahlsten, Jaakko; Jaskari, Joel; Kaski, Kimmo; Mäkitie, Antti A.; Fuller, Clifton D.; Aerts, Hugo J.W.L.; Kann, Benjamin H.; Department of Computer Science; Kaski Kimmo group; Harvard Medical School; Harvard School of Public Health; University of Texas MD Anderson Cancer Center; Department of Computer Science; University of HelsinkiImportance: Sarcopenia is an established prognostic factor in patients with head and neck squamous cell carcinoma (HNSCC); the quantification of sarcopenia assessed by imaging is typically achieved through the skeletal muscle index (SMI), which can be derived from cervical skeletal muscle segmentation and cross-sectional area. However, manual muscle segmentation is labor intensive, prone to interobserver variability, and impractical for large-scale clinical use. Objective: To develop and externally validate a fully automated image-based deep learning platform for cervical vertebral muscle segmentation and SMI calculation and evaluate associations with survival and treatment toxicity outcomes. Design, Setting, and Participants: For this prognostic study, a model development data set was curated from publicly available and deidentified data from patients with HNSCC treated at MD Anderson Cancer Center between January 1, 2003, and December 31, 2013. A total of 899 patients undergoing primary radiation for HNSCC with abdominal computed tomography scans and complete clinical information were selected. An external validation data set was retrospectively collected from patients undergoing primary radiation therapy between January 1, 1996, and December 31, 2013, at Brigham and Women's Hospital. The data analysis was performed between May 1, 2022, and March 31, 2023. Exposure: C3 vertebral skeletal muscle segmentation during radiation therapy for HNSCC. Main Outcomes and Measures: Overall survival and treatment toxicity outcomes of HNSCC. Results: The total patient cohort comprised 899 patients with HNSCC (median [range] age, 58 [24-90] years; 140 female [15.6%] and 755 male [84.0%]). Dice similarity coefficients for the validation set (n = 96) and internal test set (n = 48) were 0.90 (95% CI, 0.90-0.91) and 0.90 (95% CI, 0.89-0.91), respectively, with a mean 96.2% acceptable rate between 2 reviewers on external clinical testing (n = 377). Estimated cross-sectional area and SMI values were associated with manually annotated values (Pearson r = 0.99; P < .001) across data sets. On multivariable Cox proportional hazards regression, SMI-derived sarcopenia was associated with worse overall survival (hazard ratio, 2.05; 95% CI, 1.04-4.04; P = .04) and longer feeding tube duration (median [range], 162 [6-1477] vs 134 [15-1255] days; hazard ratio, 0.66; 95% CI, 0.48-0.89; P = .006) than no sarcopenia. Conclusions and Relevance: This prognostic study's findings show external validation of a fully automated deep learning pipeline to accurately measure sarcopenia in HNSCC and an association with important disease outcomes. The pipeline could enable the integration of sarcopenia assessment into clinical decision making for individuals with HNSCC.Item Harnessing uncertainty in radiotherapy auto-segmentation quality assurance(Elsevier, 2024-01) Wahid, Kareem A.; Sahlsten, Jaakko; Jaskari, Joel; Dohopolski, Michael J.; Kaski, Kimmo; He, Renjie; Glerean, Enrico; Kann, Benjamin H.; Mäkitie, Antti; Fuller, Clifton D.; Naser, Mohamed A.; Fuentes, David; Department of Computer Science; Department of Neuroscience and Biomedical Engineering; Kaski Kimmo group; Department of Computer Science; University of Texas Southwestern Medical Center; University of Texas MD Anderson Cancer Center; Harvard Medical School; University of HelsinkiItem Implantoitavat bioanturit(2015-05-25) Jaskari, Joel; Turunen, Markus; Sähkötekniikan korkeakoulu; Turunen, MarkusItem Machine Learning for Healthcare(Aalto University, 2022) Jaskari, Joel; Särkkä, Simo, Prof., Aalto University, Finland; Kaski, Kimmo, Prof., Aalto University, Finland; Tietotekniikan laitos; Department of Computer Science; Perustieteiden korkeakoulu; School of Science; Solin, Arno, Prof., Aalto University, Department of Computer Science, FinlandMachine learning has been recently proposed for various medical applications. Especially the deep neural network based approach has been found to achieve state-of-the-art performance in various classification tasks. However, many of these studies use simplified classification systems, for example, the referable/non-referable system in the case of diabetic retinopathy classification. Moreover, the studies that have used clinical classification systems have not considered the uncertainty of the classifiers, which is of paramount interest in the medical field. In addition, extensive analysis of automatic segmentation algorithms that includes comparison to the interobserver variability of multiple radiologists' segmentations has not yet been performed for some challenging tasks, such as the automatic segmentation of the mandibular canals. The machine learning algorithms should also be able to be trained on local hospital data, which can pose issues relating to the amount of available training data. This thesis considers machine learning for various tasks in healthcare using Finnish hospital data. Deep convolutional neural networks (CNNs) are utilized for diabetic retinopathy and macular edema classification based on clinical severity scales. In addition, approximate Bayesian deep learning approaches are systematically studied for uncertainty-aware diabetic retinopathy classification of clinical data. A connection is derived between the referral of uncertain classifications and reject option classification, and it is used to develop a novel uncertainty measure. A CNN approach will also be introduced for the segmentation of the mandibular canal in cone beam computed tomography volumes. The approach is then compared to the interobserver variability of multiple radiologists' segmentations of the canal. Lastly, this thesis will examine multiple machine learning approaches for very low birth weight neonate mortality and morbidity prediction. The results suggest that even a relatively small set of Finnish hospital data can be utilized to train deep learning classifiers for diabetic retinopathy and macular edema classification with clinical classification systems. It also turns out that approximate Bayesian neural networks and the derived novel uncertainty measure can be used to accurately estimate the uncertainty in clinical diabetic retinopathy classification. The deep learning approach is shown to set a new state-of-the-art for the mandibular canal segmentation task and it is also found to localize the canals with lower variability than the interobserver variability of four radiologists. A random forest classifier turned out to outperform other methods in neonatal mortality and morbidity prediction.Item Machine Learning Methods for Neonatal Mortality and Morbidity Classification(IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC, 2020-07-02) Jaskari, Joel; Myllarinen, Janne; Leskinen, Markus; Rad, Ali Bahrami; Hollmén, Jaakko; Andersson, Sture; Sarkka, Simo; Department of Computer Science; Department of Electrical Engineering and Automation; Kaski Kimmo group; Sensor Informatics and Medical Technology; Helsinki Institute for Information Technology (HIIT); Professorship Rousu Juho; Department of Electrical Engineering and Automation; University of HelsinkiPreterm birth is the leading cause of mortality in children under the age of five. In particular, low birth weight and low gestational age are associated with an increased risk of mortality. Preterm birth also increases the risks of several complications, which can increase the risk of death, or cause long-term morbidities with both individual and societal impacts. In this work, we use machine learning for prediction of neonatal mortality as well as neonatal morbidities of bronchopulmonary dysplasia, necrotizing enterocolitis, and retinopathy of prematurity, among very low birth weight infants. Our predictors include time series data and clinical variables collected at the neonatal intensive care unit of Children's Hospital, Helsinki University Hospital. We examine 9 different classifiers and present our main results in AUROC, similar to our previous studies, and in F1-score, which we propose for classifier selection in this study. We also investigate how the predictive performance of the classifiers evolves as the length of time series is increased, and examine the relative importance of different features using the random forest classifier, which we found to generally perform the best in all tasks. Our systematic study also involves different data preprocessing methods which can be used to improve classifier sensitivities. Our best classifier AUROC is 0.922 in the prediction of mortality, 0.899 in the prediction of bronchopulmonary dysplasia, 0.806 in the prediction of necrotizing enterocolitis, and 0.846 in the prediction of retinopathy of prematurity. Our best classifier F1-score is 0.493 in the prediction of mortality, 0.704 in the prediction of bronchopulmonary dysplasia, 0.215 in the prediction of necrotizing enterocolitis, and 0.368 in the prediction of retinopathy of prematurity.Item Muscle and adipose tissue segmentations at the third cervical vertebral level in patients with head and neck cancer(Nature Publishing Group, 2022-08-02) Wahid, Kareem A.; Olson, Brennan; Jain, Rishab; Grossberg, Aaron J.; El-Habashy, Dina; Dede, Cem; Salama, Vivian; Abobakr, Moamen; Mohamed, Abdallah S.R.; He, Renjie; Jaskari, Joel; Sahlsten, Jaakko; Kaski, Kimmo; Fuller, Clifton D.; Naser, Mohamed A.; Department of Computer Science; Kaski Kimmo group; University of Texas MD Anderson Cancer Center; Oregon Health and Science University; Menoufia University; Department of Computer ScienceThe accurate determination of sarcopenia is critical for disease management in patients with head and neck cancer (HNC). Quantitative determination of sarcopenia is currently dependent on manually-generated segmentations of skeletal muscle derived from computed tomography (CT) cross-sectional imaging. This has prompted the increasing utilization of machine learning models for automated sarcopenia determination. However, extant datasets currently do not provide the necessary manually-generated skeletal muscle segmentations at the C3 vertebral level needed for building these models. In this data descriptor, a set of 394 HNC patients were selected from The Cancer Imaging Archive, and their skeletal muscle and adipose tissue was manually segmented at the C3 vertebral level using sliceOmatic. Subsequently, using publicly disseminated Python scripts, we generated corresponding segmentations files in Neuroimaging Informatics Technology Initiative format. In addition to segmentation data, additional clinical demographic data germane to body composition analysis have been retrospectively collected for these patients. These data are a valuable resource for studying sarcopenia and body composition analysis in patients with HNC.Item Neural Network Compression for Diagnosis of Diabetic Retinopathy(2020-12-07) Keskinen, Milli; Jaskari, Joel; Sähkötekniikan korkeakoulu; Turunen, MarkusItem Reproducibility analysis of automated deep learning based localisation of mandibular canals on a temporal CBCT dataset(Nature Publishing Group, 2023-12) Järnstedt, Jorma; Sahlsten, Jaakko; Jaskari, Joel; Kaski, Kimmo; Mehtonen, Helena; Hietanen, Ari; Sundqvist, Osku; Varjonen, Vesa; Mattila, Vesa; Prapayasatok, Sangsom; Nalampang, Sakarat; Department of Computer Science; Kaski Kimmo group; Department of Computer Science; Planmeca Oy; Chiang Mai University; Tampere UniversityPreoperative radiological identification of mandibular canals is essential for maxillofacial surgery. This study demonstrates the reproducibility of a deep learning system (DLS) by evaluating its localisation performance on 165 heterogeneous cone beam computed tomography (CBCT) scans from 72 patients in comparison to an experienced radiologist’s annotations. We evaluated the performance of the DLS using the symmetric mean curve distance (SMCD), the average symmetric surface distance (ASSD), and the Dice similarity coefficient (DSC). The reproducibility of the SMCD was assessed using the within-subject coefficient of repeatability (RC). Three other experts rated the diagnostic validity twice using a 0–4 Likert scale. The reproducibility of the Likert scoring was assessed using the repeatability measure (RM). The RC of SMCD was 0.969 mm, the median (interquartile range) SMCD and ASSD were 0.643 (0.186) mm and 0.351 (0.135) mm, respectively, and the mean (standard deviation) DSC was 0.548 (0.138). The DLS performance was most affected by postoperative changes. The RM of the Likert scoring was 0.923 for the radiologist and 0.877 for the DLS. The mean (standard deviation) Likert score was 3.94 (0.27) for the radiologist and 3.84 (0.65) for the DLS. The DLS demonstrated proficient qualitative and quantitative reproducibility, temporal generalisability, and clinical validity.Item Segmentation stability of human head and neck cancer medical images for radiotherapy applications under de-identification conditions: Benchmarking data sharing and artificial intelligence use-cases(Frontiers Research Foundation, 2023) Sahlsten, Jaakko; Wahid, Kareem A.; Glerean, Enrico; Jaskari, Joel; Naser, Mohamed A.; He, Renjie; Kann, Benjamin H.; Mäkitie, Antti; Fuller, Clifton D.; Kaski, Kimmo; Department of Computer Science; Department of Neuroscience and Biomedical Engineering; Kaski Kimmo group; Department of Computer Science; University of Texas MD Anderson Cancer Center; Harvard Medical School; University of HelsinkiBackground: Demand for head and neck cancer (HNC) radiotherapy data in algorithmic development has prompted increased image dataset sharing. Medical images must comply with data protection requirements so that re-use is enabled without disclosing patient identifiers. Defacing, i.e., the removal of facial features from images, is often considered a reasonable compromise between data protection and re-usability for neuroimaging data. While defacing tools have been developed by the neuroimaging community, their acceptability for radiotherapy applications have not been explored. Therefore, this study systematically investigated the impact of available defacing algorithms on HNC organs at risk (OARs). Methods: A publicly available dataset of magnetic resonance imaging scans for 55 HNC patients with eight segmented OARs (bilateral submandibular glands, parotid glands, level II neck lymph nodes, level III neck lymph nodes) was utilized. Eight publicly available defacing algorithms were investigated: afni_refacer, DeepDefacer, defacer, fsl_deface, mask_face, mri_deface, pydeface, and quickshear. Using a subset of scans where defacing succeeded (N=29), a 5-fold cross-validation 3D U-net based OAR auto-segmentation model was utilized to perform two main experiments: 1.) comparing original and defaced data for training when evaluated on original data; 2.) using original data for training and comparing the model evaluation on original and defaced data. Models were primarily assessed using the Dice similarity coefficient (DSC). Results: Most defacing methods were unable to produce any usable images for evaluation, while mask_face, fsl_deface, and pydeface were unable to remove the face for 29%, 18%, and 24% of subjects, respectively. When using the original data for evaluation, the composite OAR DSC was statistically higher (p ≤ 0.05) for the model trained with the original data with a DSC of 0.760 compared to the mask_face, fsl_deface, and pydeface models with DSCs of 0.742, 0.736, and 0.449, respectively. Moreover, the model trained with original data had decreased performance (p ≤ 0.05) when evaluated on the defaced data with DSCs of 0.673, 0.693, and 0.406 for mask_face, fsl_deface, and pydeface, respectively. Conclusion: Defacing algorithms may have a significant impact on HNC OAR auto-segmentation model training and testing. This work highlights the need for further development of HNC-specific image anonymization methods.Item Semi-Supervised Learning in Retinal Image Analysis(2018-06-18) Jaskari, Joel; Kivinen, Jyri; Perustieteiden korkeakoulu; Kaski, KimmoThis thesis studies automated detection of diabetic retinopathy in color retinal images using deep learning. Diabetic retinopathy is a condition caused by diabetes, giving rise to complications in the back of the eye, called retina. The complications of diabetic retinopathy may cause partial or complete loss of vision, if left untreated. The screening for diabetic retinopathy in the population of diabetics is vital, however the screening is currently performed by medical professionals, only. To reduce the amount of manual labour required and to make the screening more accessible to people worldwide, automated systems for the detection of diabetic retinopathy are needed. Recently proposed deep learning systems have reached high specificity and high sensitivity for the detection of diabetic retinopathy. However, these systems use large datasets of labeled examples in the training. In this thesis we study the training of a deep learning classifier, in the detection of diabetic retinopathy, using less annotated data than the state of the art systems, in both supervised and semi-supervised learning. We develop a novel parametrization of the variational autoencoder (VAE), called class hierarchical variational autoencoder (CHVAE), which we implement in the semi-supervised setting. All models reached sensitivity and specificity recommended for a human screener. However, semi-supervised learning did not improve the classification results. The CHVAE did produce more realistic samples of retinal images than previously proposed VAE and GAN models for retinal image synthesis, however it did not produce as realistic samples as the state of the art system, and it did not generate samples with clear signs of diabetic retinopathy.
