Clinical Learning Model Research Articles

Deep learning Radiomics Based on Two-Dimensional Ultrasound for Predicting the Efficacy of Neoadjuvant Chemotherapy in Breast Cancer.

We investigate the predictive value of a comprehensive model based on preoperative ultrasound radiomics, deep learning, and clinical features for pathological complete response (pCR) after neoadjuvant chemotherapy (NAC) for the breast cancer. We enrolled 155 patients with pathologically confirmed breast cancer who underwent NAC. The patients were randomly divided into the training set and the validation set in the ratio of 7:3. The deep learning and radiomics features of pre-treatment ultrasound images were extracted, and the random forest recursive elimination algorithm and the least absolute shrinkage and selection operator were used for feature screening and DL-Score and Rad-Score construction. According to multifactorial logistic regression, independent clinical predictors, DL-Score, and Rad-Score were selected to construct the comprehensive prediction model DLRC. The performance of the model was evaluated in terms of its predictive effect, and clinical practicability. Compared to the clinical, radiomics (Rad-Score), and deep learning (DL-Score) models, the DLRC accurately predicted the pCR status, with an area under the curve (AUC) of 0.937 (95%CI: 0.895-0.970) in the training set and 0.914 (95%CI: 0.838-0.973) in the validation set. Moreover, decision curve analysis confirmed that the DLRC had the highest clinical value among all models. The comprehensive model DLRC based on ultrasound radiomics, deep learning, and clinical features can effectively and accurately predict the pCR status of breast cancer after NAC, which is conducive to assisting clinical personalized diagnosis and treatment plan.

Lab2clean: a novel algorithm for automated cleaning of retrospective clinical laboratory results data for secondary uses

BackgroundThe integrity of clinical research and machine learning models in healthcare heavily relies on the quality of underlying clinical laboratory data. However, the preprocessing of this data to ensure its reliability and accuracy remains a significant challenge due to variations in data recording and reporting standards.MethodsWe developed lab2clean, a novel algorithm aimed at automating and standardizing the cleaning of retrospective clinical laboratory results data. lab2clean was implemented as two R functions specifically designed to enhance data conformance and plausibility by standardizing result formats and validating result values. The functionality and performance of the algorithm were evaluated using two extensive electronic medical record (EMR) databases, encompassing various clinical settings.Resultslab2clean effectively reduced the variability of laboratory results and identified potentially erroneous records. Upon deployment, it demonstrated effective and fast standardization and validation of substantial laboratory data records. The evaluation highlighted significant improvements in the conformance and plausibility of lab results, confirming the algorithm’s efficacy in handling large-scale data sets.Conclusionslab2clean addresses the challenge of preprocessing and cleaning clinical laboratory data, a critical step in ensuring high-quality data for research outcomes. It offers a straightforward, efficient tool for researchers, improving the quality of clinical laboratory data, a major portion of healthcare data. Thereby, enhancing the reliability and reproducibility of clinical research outcomes and clinical machine learning models. Future developments aim to broaden its functionality and accessibility, solidifying its vital role in healthcare data management.Graphical

Clinical and dermoscopy image-based deep learning models for skin lesion diagnosis in clinical practice

Clinical and dermoscopy image-based deep learning models for skin lesion diagnosis in clinical practice

Ultrasound deep learning radiomics and clinical machine learning models to predict low nuclear grade, ER, PR, and HER2 receptor status in pure ductal carcinoma in situ.

Low nuclear grade ductal carcinoma in situ (DCIS) patients can adopt proactive management strategies to avoid unnecessary surgical resection. Different personalized treatment modalities may be selected based on the expression status of molecular markers, which is also predictive of different outcomes and risks of recurrence. DCIS ultrasound findings are mostly non mass lesions, making it difficult to determine boundaries. Currently, studies have shown that models based on deep learning radiomics (DLR) have advantages in automatic recognition of tumor contours. Machine learning models based on clinical imaging features can explain the importance of imaging features. The available ultrasound data of 349 patients with pure DCIS confirmed by surgical pathology [54 low nuclear grade, 175 positive estrogen receptor (ER+), 163 positive progesterone receptor (PR+), and 81 positive human epidermal growth factor receptor 2 (HER2+)] were collected. Radiologists extracted ultrasonographic features of DCIS lesions based on the 5th Edition of Breast Imaging Reporting and Data System (BI-RADS). Patient age and BI-RADS characteristics were used to construct clinical machine learning (CML) models. The RadImageNet pretrained network was used for extracting radiomics features and as an input for DLR modeling. For training and validation datasets, 80% and 20% of the data, respectively, were used. Logistic regression (LR), support vector machine (SVM), random forest (RF), and eXtreme Gradient Boosting (XGBoost) algorithms were performed and compared for the final classification modeling. Each task used the area under the receiver operating characteristic curve (AUC) to evaluate the effectiveness of DLR and CML models. In the training dataset, low nuclear grade, ER+, PR+, and HER2+ DCIS lesions accounted for 19.20%, 65.12%, 61.21%, and 30.19%, respectively; the validation set, they consisted of 19.30%, 62.50%, 57.14%, and 30.91%, respectively. In the DLR models we developed, the best AUC values for identifying features were 0.633 for identifying low nuclear grade, completed by the XGBoost Classifier of ResNet50; 0.618 for identifying ER, completed by the RF Classifier of InceptionV3; 0.755 for identifying PR, completed by the XGBoost Classifier of InceptionV3; and 0.713 for identifying HER2, completed by the LR Classifier of ResNet50. The CML models had better performance than DLR in predicting low nuclear grade, ER+, PR+, and HER2+ DCIS lesions. The best AUC values by classification were as follows: for low nuclear grade by RF classification, AUC: 0.719; for ER+ by XGBoost classification, AUC: 0.761; for PR+ by XGBoost classification, AUC: 0.780; and for HER2+ by RF classification, AUC: 0.723. Based on small-scale datasets, our study showed that the DLR models developed using RadImageNet pretrained network and CML models may help predict low nuclear grade, ER+, PR+, and HER2+ DCIS lesions so that patients benefit from hierarchical and personalized treatment.

Deep learning-based multi-parametric magnetic resonance imaging (mp-MRI) nomogram for predicting Ki-67 expression in rectal cancer.

To explore the value of deep learning-based multi-parametric magnetic resonance imaging (mp-MRI) nomogram in predicting the Ki-67 expression in rectal cancer. The data of 491 patients with rectal cancer from two centers were retrospectively analyzed and divided into training, internal validation, and external validation sets. They were categorized into high- and low-expression group based on postoperative pathological Ki-67 expression. Each patient's mp-MRI data were analyzed to extract and select the most relevant features of deep learning, and a deep learning model was constructed. Independent predictive risk factors were identified and incorporated into a clinical model, and the clinical and deep learning models were combined to obtain a nomogram for the prediction of Ki-67 expression. The performance characteristics of the DL-model, clinical model, and nomogram were assessed using ROCs, calibration curve, decision curve, and clinical impact curve analysis. The strongest deep learning features were extracted and screened from mp-MRI data. Two independent predictive factors, namely Magnetic Resonance Imaging T (mrT) staging and differentiation degree, were identified through clinical feature selection. Three models were constructed: a deep learning (DL)-model, a clinical model, and a nomogram. The AUCs of clinical model in the training, internal validation, and external validation set were 0.69, 0.78, and 0.67, respectively. The AUCs of the deep model and nomogram ranged from 0.88 to 0.98. The prediction performance of the deep learning model and nomogram was significantly better than the clinical model (P < 0.001). The nomogram based on deep learning can help clinicians accurately and conveniently predict the expression status of Ki-67 in rectal cancer.

Nursing Students' and Nurse Preceptors' Perspectives of an Ambulatory Dedicated Education Unit in Primary Care and General Internal Medicine.

The sustainability of prelicensure nursing clinical learning models within ambulatory care is dependent on strong partnership and the availability of ambulatory nurse preceptors. Evaluation of preceptor and student satisfaction with ambulatory clinical experiences is necessary to identify program strengths and opportunities. Nursing students' and nurse preceptors' perceptions of an ambulatory dedicated education unit (DEU) in primary care and general internal medicine practices were measured from 2021 to 2023. Students' survey responses indicated a high level of agreement for 20 (90.9%) of 22 items in 2021, 35 (100%) of 35 items in 2022, and 32 (91.4%) of 35 items in 2023. Preceptor responses demonstrated favorable agreement with all of the work-life satisfaction across the 3 years of the survey, with the exception of one question on the 2021 survey. Nursing students and staff nurse preceptors found the ambulatory DEU teaching and learning experience to be positive and satisfying. [J Nurs Educ. 2024;63(3):171-177.].

Deep Learning Radiomic Analysis of MRI Combined with Clinical Characteristics Diagnoses Placenta Accreta Spectrum and its Subtypes.

Different placenta accreta spectrum (PAS) subtypes pose varying surgical risks to the parturient. Machine learning model has the potential to diagnose PAS disorder. To develop a cascaded deep semantic-radiomic-clinical (DRC) model for diagnosing PAS and its subtypes based on T2-weighted MRI. Retrospective. 361 pregnant women (mean age: 33.10 ± 4.37 years), suspected of PAS, divided into segment training cohort (N = 40), internal training cohort (N = 139), internal testing cohort (N = 60), and external testing cohort (N = 122). Coronal T2-weighted sequence at 1.5 T and 3.0 T. Clinical characteristics such as history of uterine surgery and the presence of placenta previa, complete placenta previa and dangerous placenta previa were extracted from clinical records. The DRC model (incorporating radiomics, deep semantic features, and clinical characteristics), a cumulative radiological score method performed by radiologists, and other models (including a radiomics and clinical, the clinical, radiomics and deep learning models) were developed for PAS disorder diagnosing (existence of PAS and its subtypes). AUC, ACC, Student's t-test, the Mann-Whitney U test, chi-squared test, dice coefficient, intraclass correlation coefficients, least absolute shrinkage and selection operator regression, receiver operating characteristic curve, calibration curve with the Hosmer-Lemeshow test, decision curve analysis, DeLong test, and McNemar test. P < 0.05 indicated a significant difference. In PAS diagnosis, the DRC-1 outperformed than other models (AUC = 0.850 and 0.841 in internal and external testing cohorts, respectively). In PAS subtype classification (abnormal adherent placenta and abnormal invasive placenta), DRC-2 model performed similarly with radiologists (P = 0.773 and 0.579 in the internal testing cohort and P = 0.429 and 0.874 in the external testing cohort, respectively). The DRC model offers efficiency and high diagnostic sensitivity in diagnosis, aiding in surgical planning. 3 TECHNICAL EFFICACY: Stage 2.

Mitigating Bias in Clinical Machine Learning Models

Mitigating Bias in Clinical Machine Learning Models

Developing a Clinical Learning Model to Improve Nursing Students' Learning Outcomes

The utilization rate of the clinical learning model in Indonesia comprises the following:conferences(54.5%),bedsideteaching(54.5%),clinic-tutorials(45.4%),journal percentage (45.4%), and case presentations (45.5%). As a result, this study aimed to improve nursing students' clinical learning outcomes in their ability to effectively address and resolve problems, make informed decisions pertaining to patient cases, and interact with patients and their families. Methods:This study employed an explanatory methodology with a quantitative approach as it elucidated the cause-and-effect relationship between variables. The study sample consisted of 45 nursing students who met the research criteria and conducted clinical practice at the Padang Pariaman Regional Hospital. The study employed total sampling as its sampling technique. Results:After the clinical learning model was implemented, the mean student clinical competency score increased from 24.51 to 26.33, according to the research findings. Similarly, the average student knowledge score also improved from 74.69 to 77.42 following the implementation of the clinical learning model. The study found a significant difference between the level of clinical proficiency students attained before and after the implementation of the clinical learning model (p= 0.001). Additionally, there was a substantial disparity in students' knowledge scores before and after the application of the clinical learning model (p= 0.000). Conclusion:The clinical learning models can improve student learning outcomes and the quality of education.

Identifying best fall-related balance factors and robotic-assisted gait training attributes in 105 post-stroke patients using clinical machine learning models.

Despite the promising effects of robot-assisted gait training (RAGT) on balance and gait in post-stroke rehabilitation, the optimal predictors of fall-related balance and effective RAGT attributes remain unclear in post-stroke patients at a high risk of fall. We aimed to determine the most accurate clinical machine learning (ML) algorithm for predicting fall-related balance factors and identifying RAGT attributes. We applied five ML algorithms- logistic regression, random forest, decision tree, support vector machine (SVM), and extreme gradient boosting (XGboost)- to a dataset of 105 post-stroke patients undergoing RAGT. The variables included the Berg Balance Scale score, walking speed, steps, hip and knee active torques, functional ambulation categories, Fugl- Meyer assessment (FMA), the Korean version of the Modified Barthel Index, and fall history. The random forest algorithm excelled (receiver operating characteristic area under the curve; AUC = 0.91) in predicting balance improvement, outperforming the SVM (AUC = 0.76) and XGboost (AUC = 0.71). Key determinants identified were knee active torque, age, step count, number of RAGT sessions, FMA, and hip torque. The random forest algorithm was the best prediction model for identifying fall-related balance and RAGT determinants, highlighting the importance of key factors for successful RAGT outcome performance in fall-related balance improvement.

Advancing delirium classification: A clinical notes-based natural language processing-supported machine learning model

ObjectiveThe study of the epidemiology of delirium in hospitalized patients is challenging. We aimed to identify the presence or absence of delirium from clinical text notes using natural language processing (NLP) techniques and machine learning (ML) models. Materials and methodsWe developed a delirium predictive model using 942 clinical notes from hospitalized patients with an ICD-10 delirium hospital discharge code. Moreover, we implemented ML models using a) delirium-suggestive words from an expert-defined dictionary or b) free text in clinical notes. Both strategies considered positive and negative delirium-associated words. ResultsAt the note level, for the dictionary method, the logistic regression model achieved an area under the receiver-operating curve (AUROC) of 0.917 for positive words and 0.914 for combined positive and negative words. The areas under the precision-recall curve (AUPR) were 0.893 and 0.897, respectively. For the free-text method, the model achieved an AUROC of 0.826 and 0.830 and AUPR of 0.852 and 0.856, respectively. DiscussionNLP-based ML models accurately identified the presence of delirium in clinical notes. The dictionary-based method was superior to the free-text method. The use of negative features improved performance in both methods. ConclusionOur proposed NLP-based ML model identified delirium in clinical notes. This model could automatically screen millions of notes and facilitate the study of the epidemiology of in-hospital delirium.

Combining a deep learning model with clinical data better predicts hepatocellular carcinoma behavior following surgery

Hepatocellular carcinoma (HCC) is among the most common cancers worldwide, and tumor recurrence following liver resection or transplantation is one of the highest contributors to mortality in HCC patients after surgery. Using artificial intelligence (AI), we developed an interdisciplinary model to predict HCC recurrence and patient survival following surgery. We collected whole-slide H&E images, clinical variables, and follow-up data from 300 patients with HCC who underwent transplant and 169 patients who underwent resection at the Cleveland Clinic. A deep learning model was trained to predict recurrence-free survival (RFS) and disease-specific survival (DSS) from the H&E-stained slides. Repeated cross-validation splits were used to compute robust C-index estimates, and the results were compared to those obtained by fitting a Cox proportional hazard model using only clinical variables. While the deep learning model alone was predictive of recurrence and survival among patients in both cohorts, integrating the clinical and histologic models significantly increased the C-index in each cohort. In every subgroup analyzed, we found that a combined clinical and deep learning model better predicted post-surgical outcome in HCC patients compared to either approach independently.

Should All Pancreatic Cystic Lesions with Worrisome or High-Risk Features Be Resected? A Clinical and Radiological Machine Learning Model May Help to Answer

According to current guidelines, pancreatic cystic lesions (PCLs) with worrisome or high-risk features may have overtreatment. The purpose of this study was to build a clinical and radiological based machine-learning (ML) model to identify malignant PCLs for surgery among preoperative PCLs with worrisome or high-risk features. Clinical and radiological details of 317 pathologically confirmed PCLs with worrisome or high-risk features were retrospectively analyzed and applied to ML models including Support Vector Machine, Logistic Regression (LR), Decision Tree, Bernoulli NB, Gaussian NB, K Nearest Neighbors and Linear Discriminant Analysis. The diagnostic ability for malignancy of the optimal model with the highest diagnostic AUC in the cross-validation procedure was further evaluated in internal (n=77) and external (n=50) testing cohorts, and was compared totwo published guidelines in internal mucinous cyst cohort. Ten clinical and radiological feature-based LR model was the optimal model with the highest AUC (0.951) in the cross-validation procedure. In the internal testing cohort, LR model reached an AUC, accuracy, sensitivity, and specificity of 0.927, 0.909, 0.914, and 0.905; in the external testing cohort, LR model reached 0.948, 0.900, 0.963, and 0.826. When compared tothe European guidelines and the ACG guidelines, LR model demonstrated significantly better accuracy and specificity in identifying malignancy, while maintaining the same high sensitivity. Clinical- and radiological-based LR model can accurately identify malignant PCLs in patients with worrisome or high-risk features, possessing diagnostic performance better than the European guidelines as well as ACG guidelines.

Development of an MRI-Based Comprehensive Model Fusing Clinical, Radiomics and Deep Learning Models for Preoperative Histological Stratification in Intracranial Solitary Fibrous Tumor.

Accurate preoperative histological stratification (HS) of intracranial solitary fibrous tumors (ISFTs) can help predict patient outcomes and develop personalized treatment plans. However, the role of a comprehensive model based on clinical, radiomics and deep learning (CRDL) features in preoperative HS of ISFT remains unclear. To investigate the feasibility of a CRDL model based on magnetic resonance imaging (MRI) in preoperative HS in ISFT. Retrospective. Three hundred and ninety-eight patients from Beijing Tiantan Hospital, Capital Medical University (primary training cohort) and 49 patients from Lanzhou University Second Hospital (external validation cohort) with ISFT based on histopathological findings (237 World Health Organization [WHO] tumor grade 1 or 2, and 210 WHO tumor grade 3). 3.0 T/T1-weighted imaging (T1) by using spin echo sequence, T2-weighted imaging (T2) by using fast spin echo sequence, and T1-weighted contrast-enhanced imaging (T1C) by using two-dimensional fast spin echo sequence. Area under the receiver operating characteristic curve (AUC) was used to assess the performance of the CRDL model and a clinical model (CM) in preoperative HS in the external validation cohort. The decision curve analysis (DCA) was used to evaluate the clinical net benefit provided by the CRDL model. Cohen's kappa, intra-/inter-class correlation coefficients (ICCs), Chi-square test, Fisher's exact test, Student's t-test, AUC, DCA, calibration curves, DeLong test. A P value <0.05 was considered statistically significant. The CRDL model had significantly better discrimination ability than the CM (AUC [95% confidence interval, CI]: 0.895 [0.807-0.912] vs. 0.810 [0.745-0.874], respectively) in the external validation cohort. The CRDL model can provide a clinical net benefit for preoperative HS at a threshold probability >20%. The proposed CRDL model holds promise for preoperative HS in ISFT, which is important for predicting patient outcomes and developing personalized treatment plans. 3 TECHNICAL EFFICACY: Stage 2.

Improving the Ability of Nursing Clinical Supervisors in Implementing Targeted Clinical Learning Models (MPKBT)

Clinical learning is a process of interaction between clinical preceptors and students as learners on the practice field. The clinical guidance process is often limited to supervising nursing students during clinical practice without any targets that need to be achieved, even students repeat the same activities every day. This study aims to determine the application of a target-based clinical learning model in improving the quality of preceptors. The study used a quasi-experimental design with a one group time series design on 40 nurses from several hospitals and educational institutions who had not received clinical preceptor training. The sampling technique used quota sampling technique according to the proportion of the number of students who were practicing in their respective institutions, the sample was given an intervention in the form of training using a target-based clinical learning model consisting of theory and practice. The practice was conducted in the laboratory and in the hospital from September to December. Evaluation using a checklist instrument for assessing preceptor competence and data analyzed using the Wilcoxon test using SPSS 25.0. The result of the study, there was a significant increase in the ability of nurses to provide nursing clinic preceptors between before and after training with a significance value of 0.020, while in the first supervisory assessment significant=0.030, second supervision significant=0.040 and third supervision significant=0.050. The data shows how closely the target-based learning model affects the quality of clinical supervisors' abilities. This is important to be applied to all prospective clinical supervisors before conducting clinical guidance.

A CT-Based Clinical, Radiological and Radiomic Machine Learning Model for Predicting Malignancy of Solid Renal Tumors (UroCCR-75).

Differentiating benign from malignant renal tumors is important for patient management, and it may be improved by quantitative CT features analysis including radiomic. This study aimed to compare performances of machine learning models using bio-clinical, conventional radiologic and 3D-radiomic features for the differentiation of benign and malignant solid renal tumors using pre-operative multiphasic contrast-enhanced CT examinations. A unicentric retrospective analysis of prospectively acquired data from a national kidney cancer database was conducted between January 2016 and December 2020. Histologic findings were obtained by robotic-assisted partial nephrectomy. Lesion images were semi-automatically segmented, allowing for a 3D-radiomic features extraction in the nephrographic phase. Conventional radiologic parameters such as shape, content and enhancement were combined in the analysis. Biological and clinical features were obtained from the national database. Eight machine learning (ML) models were trained and validated using a ten-fold cross-validation. Predictive performances were evaluated comparing sensitivity, specificity, accuracy and AUC. A total of 122 patients with 132 renal lesions, including 111 renal cell carcinomas (RCCs) (111/132, 84%) and 21 benign tumors (21/132, 16%), were evaluated (58 +/- 14 years, men 74%). Unilaterality (100/111, 90% vs. 13/21, 62%; p = 0.02), necrosis (81/111, 73% vs. 8/21, 38%; p = 0.02), lower values of tumor/cortex ratio at portal time (0.61 vs. 0.74, p = 0.01) and higher variation of tumor/cortex ratio between arterial and portal times (0.22 vs. 0.05, p = 0.008) were associated with malignancy. A total of 35 radiomics features were selected, and "intensity mean value" was associated with RCCs in multivariate analysis (OR = 0.99). After ten-fold cross-validation, a C5.0Tree model was retained for its predictive performances, yielding a sensitivity of 95%, specificity of 42%, accuracy of 87% and AUC of 0.74. Our machine learning-based model combining clinical, radiologic and radiomics features from multiphasic contrast-enhanced CT scans may help differentiate benign from malignant solid renal tumors.

Using clinical and radiomic feature-based machine learning models to predict pathological complete response in patients with esophageal squamous cell carcinoma receiving neoadjuvant chemoradiation.

This study aimed to build radiomic feature-based machine learning models to predict pathological clinical response (pCR) of neoadjuvant chemoradiation therapy (nCRT) for esophageal squamous cell carcinoma (ESCC) patients. A total of 112 ESCC patients who underwent nCRT followed by surgical treatment from January 2008 to December 2018 were recruited. According to pCR status (no visible cancer cells in primary cancer lesion), patients were categorized into primary cancer lesion pCR (ppCR) group (N = 65) and non-ppCR group (N = 47). Patients were also categorized into total pCR (tpCR) group (N = 48) and non-tpCR group (N = 64) according to tpCR status (no visible cancer cells in primary cancer lesion or lymph nodes). Radiomic features of pretreatment CT images were extracted, feature selection was performed, machine learning models were trained to predict ppCR and tpCR, respectively. A total of 620 radiomic features were extracted. For ppCR prediction models, radiomic model had an area under the curve (AUC) of 0.817 (95% CI: 0.732-0.896) in the testing set; and the combination model that included rad-score and clinical features had a great predicting performance, with an AUC of 0.891 (95% CI: 0.823-0.950) in the testing set. For tpCR prediction models, radiomic model had an AUC of 0.713 (95% CI: 0.613-0.808) in the testing set; and the combination model also had a great predicting performance, with an AUC of 0.814 (95% CI: 0.728-0.881) in the testing set. This study built machine learning models for predicting ppCR and tpCR of ESCC patients with favorable predicting performance respectively, which aided treatment plan optimization. This study significantly improved the predictive value of machine learning models based on radiomic features to accurately predict response to therapy of esophageal squamous cell carcinoma patients after neoadjuvant chemoradiation therapy, providing guidance for further treatment. • Combination model that included rad-score and clinical features had a great predicting performance. • Primary tumor pCR predicting models exhibit better predicting performance compared to corresponding total pCR predicting models.

Deep learning–based radiomic nomograms for predicting Ki67 expression in prostate cancer

BackgroundTo explore the value of a multiparametric magnetic resonance imaging (MRI)-based deep learning model for the preoperative prediction of Ki67 expression in prostate cancer (PCa).MaterialsThe data of 229 patients with PCa from two centers were retrospectively analyzed and divided into training, internal validation, and external validation sets. Deep learning features were extracted and selected from each patient’s prostate multiparametric MRI (diffusion-weighted imaging, T2-weighted imaging, and contrast-enhanced T1-weighted imaging sequences) data to establish a deep radiomic signature and construct models for the preoperative prediction of Ki67 expression. Independent predictive risk factors were identified and incorporated into a clinical model, and the clinical and deep learning models were combined to obtain a joint model. The predictive performance of multiple deep-learning models was then evaluated.ResultsSeven prediction models were constructed: one clinical model, three deep learning models (the DLRS-Resnet, DLRS-Inception, and DLRS-Densenet models), and three joint models (the Nomogram-Resnet, Nomogram-Inception, and Nomogram-Densenet models). The areas under the curve (AUCs) of the clinical model in the testing, internal validation, and external validation sets were 0.794, 0.711, and 0.75, respectively. The AUCs of the deep models and joint models ranged from 0.939 to 0.993. The DeLong test revealed that the predictive performance of the deep learning models and the joint models was superior to that of the clinical model (p < 0.01). The predictive performance of the DLRS-Resnet model was inferior to that of the Nomogram-Resnet model (p < 0.01), whereas the predictive performance of the remaining deep learning models and joint models did not differ significantly.ConclusionThe multiple easy-to-use deep learning–based models for predicting Ki67 expression in PCa developed in this study can help physicians obtain more detailed prognostic data before a patient undergoes surgery.

Fully automated segmentation and radiomics feature extraction of hypopharyngeal cancer on MRI using deep learning

ObjectivesTo use convolutional neural network for fully automated segmentation and radiomics features extraction of hypopharyngeal cancer (HPC) tumor in MRI.MethodsMR images were collected from 222 HPC patients, among them 178 patients were used for training, and another 44 patients were recruited for testing. U-Net and DeepLab V3 + architectures were used for training the models. The model performance was evaluated using the dice similarity coefficient (DSC), Jaccard index, and average surface distance. The reliability of radiomics parameters of the tumor extracted by the models was assessed using intraclass correlation coefficient (ICC).ResultsThe predicted tumor volumes by DeepLab V3 + model and U-Net model were highly correlated with those delineated manually (p < 0.001). The DSC of DeepLab V3 + model was significantly higher than that of U-Net model (0.77 vs 0.75, p < 0.05), particularly in those small tumor volumes of < 10 cm3 (0.74 vs 0.70, p < 0.001). For radiomics extraction of the first-order features, both models exhibited high agreement (ICC: 0.71–0.91) with manual delineation. The radiomics extracted by DeepLab V3 + model had significantly higher ICCs than those extracted by U-Net model for 7 of 19 first-order features and for 8 of 17 shape-based features (p < 0.05).ConclusionBoth DeepLab V3 + and U-Net models produced reasonable results in automated segmentation and radiomic features extraction of HPC on MR images, whereas DeepLab V3 + had a better performance than U-Net.Clinical relevance statementThe deep learning model, DeepLab V3 + , exhibited promising performance in automated tumor segmentation and radiomics extraction for hypopharyngeal cancer on MRI. This approach holds great potential for enhancing the radiotherapy workflow and facilitating prediction of treatment outcomes.Key Points• DeepLab V3 + and U-Net models produced reasonable results in automated segmentation and radiomic features extraction of HPC on MR images.• DeepLab V3 + model was more accurate than U-Net in automated segmentation, especially on small tumors.• DeepLab V3 + exhibited higher agreement for about half of the first-order and shape-based radiomics features than U-Net.

DASS Good: Explainable Data Mining of Spatial Cohort Data.

Developing applicable clinical machine learning models is a difficult task when the data includes spatial information, for example, radiation dose distributions across adjacent organs at risk. We describe the co-design of a modeling system, DASS, to support the hybrid human-machine development and validation of predictive models for estimating long-term toxicities related to radiotherapy doses in head and neck cancer patients. Developed in collaboration with domain experts in oncology and data mining, DASS incorporates human-in-the-loop visual steering, spatial data, and explainable AI to augment domain knowledge with automatic data mining. We demonstrate DASS with the development of two practical clinical stratification models and report feedback from domain experts. Finally, we describe the design lessons learned from this collaborative experience.