Geon Oh, Yongha Gi, Jeongshim Lee, Hunjung Kim, Hong-Gyun Wu, Jong Min Park, Eunae Choi, Dongho Shin, Myonggeun Yoon, Boram Lee, Jaeman Son
{"title":"Hybrid Approach to Classifying Histological Subtypes of Non-small Cell Lung Cancer (NSCLC): Combining Radiomics and Deep Learning Features from CT Images.","authors":"Geon Oh, Yongha Gi, Jeongshim Lee, Hunjung Kim, Hong-Gyun Wu, Jong Min Park, Eunae Choi, Dongho Shin, Myonggeun Yoon, Boram Lee, Jaeman Son","doi":"10.1007/s10278-025-01442-5","DOIUrl":"https://doi.org/10.1007/s10278-025-01442-5","url":null,"abstract":"<p><p>This study aimed to develop a hybrid model combining radiomics and deep learning features derived from computed tomography (CT) images to classify histological subtypes of non-small cell lung cancer (NSCLC). We analyzed CT images and radiomics features from 235 patients with NSCLC, including 110 with adenocarcinoma (ADC) and 112 with squamous cell carcinoma (SCC). The dataset was split into a training set (75%) and a test set (25%). External validation was conducted using the NSCLC-Radiomics database, comprising 24 patients each with ADC and SCC. A total of 1409 radiomics and 8192 deep features underwent principal component analysis (PCA) and ℓ2,1-norm minimization for feature reduction and selection. The optimal feature sets for classification included 27 radiomics features, 20 deep features, and 55 combined features (30 deep and 25 radiomics). The average area under the receiver operating characteristic curve (AUC) for radiomics, deep, and combined features were 0.6568, 0.6689, and 0.7209, respectively, across the internal and external test sets. Corresponding average accuracies were 0.6013, 0.6376, and 0.6564. The combined model demonstrated superior performance in classifying NSCLC subtypes, achieving higher AUC and accuracy in both test datasets. These results suggest that the proposed hybrid approach could enhance the accuracy and reliability of NSCLC subtype classification.</p>","PeriodicalId":516858,"journal":{"name":"Journal of imaging informatics in medicine","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-02-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143426748","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}
Ji Yeong An, Janie M Lee, Myoung-Jin Jang, Su Min Ha, Jung Min Chang
{"title":"Application of a Commercial Artificial Intelligence Software in Unilateral Mammography: Simulating Total Mastectomy Scenarios.","authors":"Ji Yeong An, Janie M Lee, Myoung-Jin Jang, Su Min Ha, Jung Min Chang","doi":"10.1007/s10278-025-01432-7","DOIUrl":"https://doi.org/10.1007/s10278-025-01432-7","url":null,"abstract":"<p><p>This study was to evaluate the performance of commercially available artificial intelligence (AI) software in unilateral mammograms simulating postmastectomy surveillance compared with AI software used in bilateral mammograms from the same women serving as controls. A retrospective database search identified consecutive women who underwent breast cancer surgery between January 2021 and December 2021. AI software was applied to the mammogram immediately preceding breast cancer diagnosis in two modes: bilateral (the standard bilateral mammography dataset) and unilateral analyses (each breast's craniocaudal and mediolateral oblique views), and their outputs were reviewed. The sensitivity, specificity, and number of marks per breast were compared between the bilateral and unilateral analyses with -5% non-inferiority margin for the difference in sensitivity and specificity between the two modes. A total of 694 women (mean age, 55.2 ± 10.8 years) with unilateral or bilateral breast cancer contributed mammograms for analysis; each breast was then separately evaluated in the unilateral postmastectomy simulation (n = 1388), of which 730 had breast cancer (52.6%) (mean invasive size = 1.5 cm) and compared with bilateral mammography analysis. The sensitivity of unilateral analysis was not inferior to that of bilateral analysis (78.6% vs. 76.7%), with a difference of 1.9%. The specificity of unilateral analysis was inferior to that in the bilateral analysis (81.5% vs. 91.9%), with a difference of -10.5% being lower than the non-inferiority margin. The average number of AI marks per breast was 0.94 (unilateral [1298/1388] and bilateral analyses [1306/1388], respectively). AI software performance in simulated unilateral mammography analysis demonstrated non-inferior sensitivity and inferior specificity compared to bilateral mammography.</p>","PeriodicalId":516858,"journal":{"name":"Journal of imaging informatics in medicine","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-02-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143426621","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}
{"title":"Impact of Combined Deep Learning Image Reconstruction and Metal Artifact Reduction Algorithm on CT Image Quality in Different Scanning Conditions for Maxillofacial Region with Metal Implants: A Phantom Study.","authors":"Gongxin Yang, Haowei Wang, Ling Liu, Qifan Ma, Huimin Shi, Ying Yuan","doi":"10.1007/s10278-024-01287-4","DOIUrl":"https://doi.org/10.1007/s10278-024-01287-4","url":null,"abstract":"<p><p>This study aims to investigate the impact of combining deep learning image reconstruction (DLIR) and metal artifacts reduction (MAR) algorithms on the quality of CT images with metal implants under different scanning conditions. Four images of the maxillofacial region in pigs were taken using different metal implants for evaluation. The scans were conducted at three different dose levels (CTDIvol: 20/10/5 mGy). The images were reconstructed using three different methods: filtered back projection (FBP), adaptive statistical iterative reconstruction with Veo at a 50% level (AV50), and DLIR at three levels (low, medium, and high). Regions of interest (ROIs) were identified in various tissues (near/far/reference fat, muscle, bone) both with and without metal implants and artifacts. Parameters such as standard deviation (SD), signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), and metal artifact index (MAI) were calculated. Additionally, two experienced radiologists evaluated the subjective image quality (IQ) using a 5-point Likert scale. (1) Both observers rated MAR generated significantly lower artifact scores than non-MAR in all types of tissues (P < 0.01), except for the far shadow and bloom in bone (phantoms 1, 3, 4) and the far bloom in muscle (phantom 3) without significant differences (P = 1.0). (2) Under the same scanning condition, DLIR at three levels produced a smaller SD than those of FBP and AV50 (P < 0.05). (3) Compared to FBP and AV50, DLIR denoted a better reduction of MAI and improvement of SNR and CNR (P < 0.05) for most tissues between the four phantoms. (4) Subjective overall IQ was superior with the increasement of DLIR level (P < 0.05) and both observers agreed that DLIR produced better artifact reductions compared with FBP and AV50. The combination of DLIR and MAR algorithms can enhance image quality, significantly reduce metal artifacts, and offer high clinical value.</p>","PeriodicalId":516858,"journal":{"name":"Journal of imaging informatics in medicine","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-02-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143426753","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}
{"title":"Ischemic Stroke Lesion Core Segmentation from CT Perfusion Scans Using Attention ResUnet Deep Learning.","authors":"Omar Ibrahim Alirr","doi":"10.1007/s10278-025-01407-8","DOIUrl":"https://doi.org/10.1007/s10278-025-01407-8","url":null,"abstract":"<p><p>Accurate segmentation of ischemic stroke lesions is crucial for refining diagnosis, prognosis, and treatment planning. Manual identification is time-consuming and challenging, especially in urgent clinical scenarios. This paper presents an innovative deep learning-based system for automated segmentation of ischemic stroke lesions from Computed Tomography Perfusion (CTP) datasets. This paper introduces a deep learning-based system designed to segment ischemic stroke lesions from Computed Tomography Perfusion (CTP) datasets. The proposed approach integrates Edge Enhancing Diffusion (EED) filtering as a preprocessing step, acting as a form of hard attention to emphasize affected regions. Besides the Attention ResUnet (AttResUnet) architecture with a modified decoder path, incorporating spatial and channel attention mechanisms to capture long-range dependencies. The system was evaluated using the ISLES challenge 2018 dataset with a fivefold cross-validation approach. The proposed framework achieved a noteworthy average Dice Similarity Coefficient (DSC) score of 59%. This performance underscores the effectiveness of combining EED filtering with attention mechanisms in the AttResUnet architecture for accurate stroke lesion segmentation. The fold-wise analysis revealed consistent performance across different data subsets, with slight variations highlighting the model's generalizability. The proposed approach offers a reliable and generalizable tool for automated ischemic stroke lesion segmentation, potentially improving efficiency and accuracy in clinical settings.</p>","PeriodicalId":516858,"journal":{"name":"Journal of imaging informatics in medicine","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-02-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143426763","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}
Mohammed A AboArab, Vassiliki T Potsika, Andrzej Skalski, Maciej Stanuch, George Gkois, Igor Koncar, David Matejevic, Alexis Theodorou, Sylvia Vagena, Fragiska Sigala, Dimitrios I Fotiadis
{"title":"DECODE-3DViz: Efficient WebGL-Based High-Fidelity Visualization of Large-Scale Images using Level of Detail and Data Chunk Streaming.","authors":"Mohammed A AboArab, Vassiliki T Potsika, Andrzej Skalski, Maciej Stanuch, George Gkois, Igor Koncar, David Matejevic, Alexis Theodorou, Sylvia Vagena, Fragiska Sigala, Dimitrios I Fotiadis","doi":"10.1007/s10278-025-01430-9","DOIUrl":"https://doi.org/10.1007/s10278-025-01430-9","url":null,"abstract":"<p><p>The DECODE-3DViz pipeline represents a major advancement in the web-based visualization of large-scale medical imaging data, particularly for peripheral artery computed tomography images. This research addresses the critical challenges of rendering high-resolution volumetric datasets via WebGL technology. By integrating progressive chunk streaming and level of detail (LOD) algorithms, DECODE-3DViz optimizes the rendering process for real-time interaction and high-fidelity visualization. The system efficiently manages WebGL texture size constraints and browser memory limitations, ensuring smooth performance even with extensive datasets. A comparative evaluation against state-of-the-art visualization tools demonstrates DECODE-3DViz's superior performance, achieving up to a 98% reduction in rendering time compared with that of competitors and maintaining a high frame rate of up to 144 FPS. Furthermore, the system exhibits exceptional GPU memory efficiency, utilizing as little as 2.6 MB on desktops, which is significantly less than the over 100 MB required by other tools. User feedback, collected through a comprehensive questionnaire, revealed high satisfaction with the tool's performance, particularly in areas such as structure definition and diagnostic capability, with an average score of 4.3 out of 5. These enhancements enable detailed and accurate visualizations of the peripheral vasculature, improving diagnostic accuracy and supporting better clinical outcomes. The DECODE-3DViz tool is open source and can be accessed at https://github.com/mohammed-abo-arab/3D_WebGL_VolumeRendering.git .</p>","PeriodicalId":516858,"journal":{"name":"Journal of imaging informatics in medicine","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-02-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143426634","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}
{"title":"Estimating the Amount of Air Inside the Stomach for Detecting Cancers on Gastric Radiographs Using Artificial Intelligence: an Observational, Cross-sectional Study.","authors":"Chiharu Kai, Takahiro Irie, Yuuki Kobayashi, Hideaki Tamori, Satoshi Kondo, Akifumi Yoshida, Yuta Hirono, Ikumi Sato, Kunihiko Oochi, Satoshi Kasai","doi":"10.1007/s10278-025-01441-6","DOIUrl":"https://doi.org/10.1007/s10278-025-01441-6","url":null,"abstract":"<p><p>Gastric radiography is an important tool for early detection of cancer. During gastric radiography, the stomach is monitored using barium and effervescent granules. However, stomach compression and physiological phenomena during the examination can cause air to escape the stomach. When the stomach contracts, physicians cannot accurately observe its condition, which may result in missed lesions. Notably, no research using artificial intelligence (AI) has explored the use of gastric radiography to estimate the amount of air in the stomach. Therefore, this study aimed to develop an AI system to estimate the amount of air inside the stomach using gastric radiographs. In this observational, cross-sectional study, we collected data from 300 cases who underwent medical screening and estimated the images with poor stomach air volume. We used pre-trained models of vision transformer (ViT) and convolutional neural network (CNN). Instead of retraining, dimensionality reduction was performed on the output features using principal component analysis, and LightGBM performed discriminative processing. The combination of ViT and CNN resulted in the highest accuracy (F-value 0.792, accuracy 0.943, sensitivity 0.738, specificity 0.978). High accuracy was maintained in the prone position, where air inside the stomach could be easily released. Combining ViT and CNN from gastric radiographs accurately identified cases of poor stomach air volume. The system was highly accurate in the prone position and proved clinically useful. The developed AI can be used to provide high-quality images to physicians and to prevent missed lesions.</p>","PeriodicalId":516858,"journal":{"name":"Journal of imaging informatics in medicine","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-02-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143426743","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}
Sima Soltanpour, Arnold Chang, Dan Madularu, Praveen Kulkarni, Craig Ferris, Chris Joslin
{"title":"3D Wasserstein Generative Adversarial Network with Dense U-Net-Based Discriminator for Preclinical fMRI Denoising.","authors":"Sima Soltanpour, Arnold Chang, Dan Madularu, Praveen Kulkarni, Craig Ferris, Chris Joslin","doi":"10.1007/s10278-025-01434-5","DOIUrl":"https://doi.org/10.1007/s10278-025-01434-5","url":null,"abstract":"<p><p>Functional magnetic resonance imaging (fMRI) is extensively used in clinical and preclinical settings to study brain function; however, fMRI data is inherently noisy due to physiological processes, hardware, and external noise. Denoising is one of the main preprocessing steps in any fMRI analysis pipeline. This process is challenging in preclinical data in comparison to clinical data due to variations in brain geometry, image resolution, and low signal-to-noise ratios. In this paper, we propose a structure-preserved algorithm based on a 3D Wasserstein generative adversarial network with a 3D dense U-net-based discriminator called 3D U-WGAN. We apply a 4D data configuration to effectively denoise temporal and spatial information in analyzing preclinical fMRI data. GAN-based denoising methods often utilize a discriminator to identify significant differences between denoised and noise-free images, focusing on global or local features. To refine the fMRI denoising model, our method employs a 3D dense U-Net discriminator to learn both global and local distinctions. To tackle potential oversmoothing, we introduce an adversarial loss and enhance perceptual similarity by measuring feature space distances. Experiments illustrate that 3D U-WGAN significantly improves image quality in resting-state and task preclinical fMRI data, enhancing signal-to-noise ratio without introducing excessive structural changes in existing methods. The proposed method outperforms state-of-the-art methods when applied to simulated and real data in a fMRI analysis pipeline.</p>","PeriodicalId":516858,"journal":{"name":"Journal of imaging informatics in medicine","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-02-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143412209","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}
Francesca Pia Villani, Maria Chiara Fiorentino, Lorenzo Federici, Cesare Piazza, Emanuele Frontoni, Alberto Paderno, Sara Moccia
{"title":"A Deep-Learning Approach for Vocal Fold Pose Estimation in Videoendoscopy.","authors":"Francesca Pia Villani, Maria Chiara Fiorentino, Lorenzo Federici, Cesare Piazza, Emanuele Frontoni, Alberto Paderno, Sara Moccia","doi":"10.1007/s10278-025-01431-8","DOIUrl":"https://doi.org/10.1007/s10278-025-01431-8","url":null,"abstract":"<p><p>Accurate vocal fold (VF) pose estimation is crucial for diagnosing larynx diseases that can eventually lead to VF paralysis. The videoendoscopic examination is used to assess VF motility, usually estimating the change in the anterior glottic angle (AGA). This is a subjective and time-consuming procedure requiring extensive expertise. This research proposes a deep learning framework to estimate VF pose from laryngoscopy frames acquired in the actual clinical practice. The framework performs heatmap regression relying on three anatomically relevant keypoints as a prior for AGA computation, which is estimated from the coordinates of the predicted points. The assessment of the proposed framework is performed using a newly collected dataset of 471 laryngoscopy frames from 124 patients, 28 of whom with cancer. The framework was tested in various configurations and compared with other state-of-the-art approaches (direct keypoints regression and glottal segmentation) for both pose estimation, and AGA evaluation. The proposed framework obtained the lowest root mean square error (RMSE) computed on all the keypoints (5.09, 6.56, and 6.40 pixels, respectively) among all the models tested for VF pose estimation. Also for the AGA evaluation, heatmap regression reached the lowest mean average error (MAE) ( <math><mrow><mn>5</mn> <mo>.</mo> <msup><mn>87</mn> <mo>∘</mo></msup> </mrow> </math> ). Results show that relying on keypoints heatmap regression allows to perform VF pose estimation with a small error, overcoming drawbacks of state-of-the-art algorithms, especially in challenging images such as pathologic subjects, presence of noise, and occlusion.</p>","PeriodicalId":516858,"journal":{"name":"Journal of imaging informatics in medicine","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-02-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143412210","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}
Jakob Leonhardi, Maike Niebur, Anne-Kathrin Höhn, Sebastian Ebel, Manuel Florian Struck, Hans-Michael Tautenhahn, Daniel Seehofer, Silke Zimmermann, Timm Denecke, Hans-Jonas Meyer
{"title":"Impact of MRI Texture Analysis on Complication Rate in MRI-Guided Liver Biopsies.","authors":"Jakob Leonhardi, Maike Niebur, Anne-Kathrin Höhn, Sebastian Ebel, Manuel Florian Struck, Hans-Michael Tautenhahn, Daniel Seehofer, Silke Zimmermann, Timm Denecke, Hans-Jonas Meyer","doi":"10.1007/s10278-025-01439-0","DOIUrl":"https://doi.org/10.1007/s10278-025-01439-0","url":null,"abstract":"<p><p>Magnetic resonance imaging (MRI)-derived texture features are quantitative imaging parameters that may have valuable associations with clinical aspects. Their prognostic ability in patients undergoing percutaneous MRI-guided liver biopsy to identify associations with post-interventional bleeding complications and biopsy success rate has not been sufficiently investigated. The patient sample consisted 79 patients (32 females, 40.5%) with a mean age of 58.7 ± 12.4 years. Clinical parameters evaluated included comorbidities, pre-existing liver disease, known cancer diagnosis, and hemostaseological parameters. Several puncture-related parameters such as biopsy angle, distance of needle entry to capsule, and lesion were analyzed. MRI texture features of the target lesion were extracted from the planning sequence of the MRI-guided liver biopsy. Mann-Whitney U test and Fisher's exact test were used for group comparison; multivariate regression model was used for outcome prediction. Overall, the diagnostic outcome of biopsy was malignant in 38 cases (48.1%) and benign in 32 cases (40.5%). A total of 11 patients (13.9%) had post-interventional bleeding, while nine patients (11.4%) had a negative biopsy result. Several texture features were statistically significantly different between patients with and without hemorrhage. The texture feature GrVariance (1.37 ± 0.78 vs. 0.80 ± 0.35, p = 0.007) reached the highest statistical significance. Regarding unsuccessful biopsy results, S(1,1)DifEntrp (0.80 ± 0.10 vs. 0.89 ± 0.12, p = 0.022) and S(0,4)DifEntrp (1.14 ± 0.10 vs. 1.22 ± 0.11, p = 0.021) reached statistical significance between groups. Several MRI texture features of the target lesion were associated with bleeding complications or negative biopsy after MRI-guided percutaneous liver biopsy. This could be used to identify at-risk patients at the beginning of the procedure and should be further analyzed.</p>","PeriodicalId":516858,"journal":{"name":"Journal of imaging informatics in medicine","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-02-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143401146","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}
Fernanda Veloso Pereira, Davi Ferreira, Heraldo Garmes, Denise Engelbrecht Zantut-Wittmann, Fabio Rogério, Mateus Dal Fabbro, Cleiton Formentin, Carlos Henrique Quartucci Forster, Fabiano Reis
{"title":"Machine Learning Prediction of Pituitary Macroadenoma Consistency: Utilizing Demographic Data and Brain MRI Parameters.","authors":"Fernanda Veloso Pereira, Davi Ferreira, Heraldo Garmes, Denise Engelbrecht Zantut-Wittmann, Fabio Rogério, Mateus Dal Fabbro, Cleiton Formentin, Carlos Henrique Quartucci Forster, Fabiano Reis","doi":"10.1007/s10278-025-01417-6","DOIUrl":"https://doi.org/10.1007/s10278-025-01417-6","url":null,"abstract":"<p><p>Consistency of pituitary macroadenomas is a key determinant in surgical outcomes, with non-soft consistency linked to more complications and incomplete resections. This study aimed to develop a machine learning model to predict the consistency of pituitary macroadenomas to improve surgical planning and outcomes. A retrospective study of patients with pituitary macroadenomas was conducted. Data included brain magnetic resonance imaging findings (diameter and apparent diffusion coefficient), patient demographics (age and sex), and tumor consistency. Seventy patients were evaluated, 59 with soft consistency and 11 with non-soft consistency. The support vector machine (SVM) was the best model with ROC AUC score of 83.3% [95% CI 65.8, 97.6], AP AUC of 69.8% [95% CI 41.3, 91.1], sensitivity of 73.1% [95% CI 44.4, 100], specificity of 89.8% [95% CI 82, 96.7], F1 score of 0.63 [95% CI 0.36, 0.83], and Matthews correlation coefficient score of 0.57 [95% CI 0.29, 0.79]. These findings indicate a significant improvement over random classification, as confirmed by a permutation test (p < 0.05). Additionally, the model had a 67.4% probability of outperforming the second-best model in cross-validation, as determined through Bayesian analysis, and demonstrated statistical significance (p < 0.05) compared to non-ensemble models. Using explainability heuristics, both 2D and 3D probability maps highlighted areas with a higher probability of non-soft consistency. The attributes most influential in the correct classification by our best model were male sex and age ≤ 42.25 years. Despite some limitations, the SVM model showed promise in predicting tumor consistency, which could aid in surgical planning. To address concerns about generalizability, we have created an open-access repository to promote future external validation studies and collaboration with other research centers, with the goal of enhancing model prediction through transfer learning.</p>","PeriodicalId":516858,"journal":{"name":"Journal of imaging informatics in medicine","volume":" ","pages":""},"PeriodicalIF":0.0,"publicationDate":"2025-02-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143371419","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}