Artificial Intelligence-Based Clinical Decision Support Systems Using Advanced Medical Imaging and Radiomics

Radiomics for clinical decision support in radiation oncology

Radiomics is a promising tool for the development of quantitative biomarkers to support clinical decision-making. It has been shown to improve the prediction of response to treatment and outcome in different settings, particularly in the field of radiation oncology by optimising the dose delivery solutions and reducing the rate of radiation-induced side effects, leading to a fully personalised approach. Despite the promising results offered by radiomics at each of these stages, standardised methodologies, reproducibility and interpretability of results are still lacking, limiting the potential clinical impact of these tools. In this review, we briefly describe the principles of radiomics and the most relevant applications of radiomics at each stage of cancer management in the framework of radiation oncology. Furthermore, the integration of radiomics into clinical decision support systems is analysed, defining the challenges and offering possible solutions for translating radiomics into a clinically applicable tool.

Training and assessing convolutional neural network performance in automatic vascular segmentation using Ga-68 DOTATATE PET/CT

To evaluate a convolutional neural network's performance (nnU-Net) in the assessment of vascular contours, calcification and PET tracer activity using Ga-68 DOTATATE PET/CT. Patients who underwent Ga-68 DOTATATE PET/CT imaging over a 12-month period for neuroendocrine investigation were included. Manual cardiac and aortic segmentations were performed by an experienced observer. Scans were randomly allocated in ratio 64:16:20 for training, validation and testing of the nnU-Net model. PET tracer uptake and calcium scoring were compared between segmentation methods and different observers. 116 patients (53.5% female) with a median age of 64.5 years (range 23-79) were included. There were strong, positive correlations between all segmentations (mostly r > 0.98). There were no significant differences between manual and AI segmentation of SUVmean for global cardiac (mean ± SD 0.71 ± 0.22 vs. 0.71 ± 0.22; mean diff 0.001 ± 0.008, p > 0.05), ascending aorta (mean ± SD 0.44 ± 0.14 vs. 0.44 ± 0.14; mean diff 0.002 ± 0.01, p > 0.05), aortic arch (mean ± SD 0.44 ± 0.10 vs. 0.43 ± 0.10; mean diff 0.008 ± 0.16, p > 0.05) and descending aorta (mean ± SD < 0.001; 0.58 ± 0.12 vs. 0.57 ± 0.12; mean diff 0.01 ± 0.03, p > 0.05) contours. There was excellent agreement between the majority of manual and AI segmentation measures (r ≥ 0.80) and in all vascular contour calcium scores. Compared with the manual segmentation approach, the CNN required a significantly lower workflow time. AI segmentation of vascular contours using nnU-Net resulted in very similar measures of PET tracer uptake and vascular calcification when compared to an experienced observer and significantly reduced workflow time.

AI implementation in the UK landscape: Knowledge of AI governance, perceived challenges and opportunities, and ways forward for radiographers

INTRODUCTION

Despite the rapid increase of AI-enabled applications deployed in clinical practice, many challenges exist around AI implementation, including the clarity of governance frameworks, usability of validation of AI models, and customisation of training for radiographers. This study aimed to explore the perceptions of diagnostic and therapeutic radiographers, with existing theoretical and/or practical knowledge of AI, on issues of relevance to the field, such as AI implementation, including knowledge of AI governance and procurement, perceptions about enablers and challenges and future priorities for AI adoption.

METHODS

An online survey was designed and distributed to UK-based qualified radiographers who work in medical imaging and/or radiotherapy and have some previous theoretical and/or practical knowledge of working with AI. Participants were recruited through the researchers' professional networks on social media with support from the AI advisory group of the Society and College of Radiographers. Survey questions related to AI training/education, knowledge of AI governance frameworks, data privacy procedures, AI implementation considerations, and priorities for AI adoption. Descriptive statistics were employed to analyse the data, and chi-square tests were used to explore significant relationships between variables.

RESULTS

In total, 88 valid responses were received. Most radiographers (56.6 %) had not received any AI-related training. Also, although approximately 63 % of them used an evaluation framework to assess AI models' performance before implementation, many (36.9 %) were still unsure about suitable evaluation methods. Radiographers requested clearer guidance on AI governance, ample time to implement AI in their practice safely, adequate funding, effective leadership, and targeted support from AI champions. AI training, robust governance frameworks, and patient and public involvement were seen as priorities for the successful implementation of AI by radiographers.

CONCLUSION

AI implementation is progressing within radiography, but without customised training, clearer governance, key stakeholder engagement and suitable new roles created, it will be hard to harness its benefits and minimise related risks.

IMPLICATIONS FOR PRACTICE

The results of this study highlight some of the priorities and challenges for radiographers in relation to AI adoption, namely the need for developing robust AI governance frameworks and providing optimal AI training.

Artificial intelligence ethics in precision oncology: balancing advancements in technology with patient privacy and autonomy

Precision oncology is a rapidly evolving field that uses advanced technologies to deliver personalized cancer care based on a patient's unique genetic and clinical profile. The use of artificial intelligence (AI) in precision oncology has shown great potential to improve diagnosis, treatment planning, and treatment outcomes. However, the integration of AI in precision oncology also raises important ethical considerations related to patient privacy, autonomy, and protection from bias. In this opinion paper, an overview is provided of previous studies that have explored the use of AI in precision oncology and the ethical considerations associated with this technology. The conclusions of these studies are compared, and the importance of approaching the use of AI in precision oncology with caution is emphasized. It is stressed that patient privacy, autonomy, and protection from bias should be made central to the development and use of AI in precision oncology. Clear guidelines and regulations must be established to ensure that AI is used ethically and for the benefit of patients. The use of AI in precision oncology has the potential to revolutionize cancer care, but it should be ensured that it striked a balance between advancements in technology and ethical considerations. In conclusion, the use of AI in precision oncology is a promising development that has the potential to improve cancer outcomes. However, ethical considerations related to patient privacy, autonomy, and protection from bias must be central to the development and use of AI in precision oncology.

MISM: A Medical Image Segmentation Metric for Evaluation of Weak Labeled Data

Performance measures are an important tool for assessing and comparing different medical image segmentation algorithms. Unfortunately, the current measures have their weaknesses when it comes to assessing certain edge cases. These limitations arise when images with a very small region of interest or without a region of interest at all are assessed. As a solution to these limitations, we propose a new medical image segmentation metric: MISm. This metric is a composition of the Dice similarity coefficient and the weighted specificity. MISm was investigated for definition gaps, an appropriate scoring gradient, and different weighting coefficients used to propose a constant value. Furthermore, an evaluation was performed by comparing the popular metrics in the medical image segmentation and MISm using images of magnet resonance tomography from several fictitious prediction scenarios. Our analysis shows that MISm can be applied in a general way and thus also covers the mentioned edge cases, which are not covered by other metrics, in a reasonable way. In order to allow easy access to MISm and therefore widespread application in the community, as well as reproducibility of experimental results, we included MISm in the publicly available evaluation framework MISeval.

Prevalence of Thalassemia in the Vietnamese Population and Building a Clinical Decision Support System for Prenatal Screening for Thalassemia

Introduction

The prevalence of thalassemia among the Vietnamese population was studied, and clinical decision support systems for prenatal screening of thalassemia were created. The aim of this report was to investigate the prevalence of thalassemia in the Vietnamese population, building a clinical decision support system for prenatal screening for thalassemia.

Methods

A cross-sectional study was conducted on pregnant women and their husbands visiting the Vietnam National Hospital of Obstetrics and Gynecology from October 2020 to December 2021. A total of 10112 medical records of first-time pregnant women and their husbands were collected.

Results

A clinical decision support system was built, including 2 different types of systems for prenatal screening for thalassemia (an expert system and 4 AI-based CDSS). One thousand nine hundred ninety-two cases were used to train and test machine learning models, while 1555 cases were used for specialized expert system evaluation. There were ten key variables for AI-based CDSS for machine learning. The four most important features in thalassemia screening were identified. The accuracy of the expert system and AI-based CDSS was compared. The rate of patients with Alpha thalassemia is 10.73% (1085 patients), the rate of patients with beta-thalassemia is 2.24% (227 patients), and 0.29% (29 patients) of patients carry both alpha-thalassemia and beta-thalassemia gene mutations. The expert system showed an accuracy of 98.45%. Among the AI-based CDSS developed, the multilayer perceptron (MLP) model was the most stable regardless of the training database (accuracy of 98,5% using all features and 97% using only the four most important features).

Conclusions

When comparing the expert system with the AI-based CDSS, the accuracy of the expert system and AI-based models was comparable. The developed expert system for prenatal thalassemia screening showed high accuracy. AI-based CDSS showed satisfactory results. Further development of such systems is promising with a view to their introduction into clinical practice.

Conformity assessment of a computer vision-based posture analysis system for the screening of postural deformation

Background

This study evaluates the conformity of using a computer vision-based posture analysis system as a screening assessment for postural deformity detection in the spine that is easily applicable to clinical practice.

Methods

One hundred forty participants were enrolled for screening of the postural deformation. Factors that determine the presence or absence of spinal deformation, such as shoulder height difference (SHD), pelvic height difference (PHD), and leg length mismatch (LLD), were used as parameters for the clinical decision support system (CDSS) using a commercial computer vision-based posture analysis system. For conformity analysis, the probability of postural deformation provided by CDSS, the Cobb angle, the PHD, and the SHD was compared and analyzed between the system and radiographic parameters. A principal component analysis (PCA) of the CDSS and correlation analysis were conducted.

Results

The Cobb angles of the 140 participants ranged from 0° to 61°, with an average of 6.16° ± 8.50°. The postural deformation of CDSS showed 94% conformity correlated with radiographic assessment. The conformity assessment results were more accurate in the participants of postural deformation with normal (0–9°) and mild (10–25°) ranges of scoliosis. The referenced SHD and the SHD of the CDSS showed statistical significance (p < 0.001) on a paired t-test. SHD and PHD for PCA were the predominant factors (PC1 SHD for 79.97%, PC2 PHD for 19.86%).

Conclusion

The CDSS showed 94% conformity for the screening of postural spinal deformity. The main factors determining diagnostic suitability were two main variables: SHD and PHD. In conclusion, a computer vision-based posture analysis system can be utilized as a safe, efficient, and convenient CDSS for early diagnosis of spinal posture deformation, including scoliosis.

Towards a guideline for evaluation metrics in medical image segmentation

In the last decade, research on artificial intelligence has seen rapid growth with deep learning models, especially in the field of medical image segmentation. Various studies demonstrated that these models have powerful prediction capabilities and achieved similar results as clinicians. However, recent studies revealed that the evaluation in image segmentation studies lacks reliable model performance assessment and showed statistical bias by incorrect metric implementation or usage. Thus, this work provides an overview and interpretation guide on the following metrics for medical image segmentation evaluation in binary as well as multi-class problems: Dice similarity coefficient, Jaccard, Sensitivity, Specificity, Rand index, ROC curves, Cohen's Kappa, and Hausdorff distance. Furthermore, common issues like class imbalance and statistical as well as interpretation biases in evaluation are discussed. As a summary, we propose a guideline for standardized medical image segmentation evaluation to improve evaluation quality, reproducibility, and comparability in the research field.

Detecting and Extracting Brain Hemorrhages from CT Images Using Generative Convolutional Imaging Scheme

Purpose

The need for computerized medical assistance for accurate detection of brain hemorrhage from Computer Tomography (CT) images is more mandatory than conventional clinical tests. Recent technologies and advanced computerized algorithms follow Artificial Intelligence (AI), Machine Learning (ML), and Deep Learning (DL) techniques to improve medical diagnosis platforms. This technology is making the diagnosis practice of brain issues easier for medical practitioners to analyze and identify diseases with an assured degree of precision and performance.

Methods

As the existing CT image analysis models use standard procedures to detect hemorrhages, the need for DL-based data analysis is essential to provide more accurate results. Generally, the existing techniques are limited with image training efficiency, image filtering procedures, and runtime system tuning modules. On the scope, this work develops a DL-based automated analysis of CT scan slices to find various levels of brain hemorrhages. Notably, this proposed system integrates Convolutional Neural Network (CNN) and Generative Adversarial Network (GAN) architectures as Integrated Generative Adversarial-Convolutional Imaging Model (IGACM) for extracting the CT image features for detecting brain hemorrhages.

Results

This system produces good results and takes lesser training time than existing techniques. This proposed system effectively works over CT images and classifies the abnormalities with more accuracy than current techniques. The experiments and results deliver the optimal detection of hemorrhages with better accuracy. It shows that the proposed system works with 5% to 10% of the better performance compared to other diagnostic techniques.

Conclusion

The complex nature of CT images leads to noncorrelated feature complexities in diagnosis models. Considering the issue, the proposed system used GAN-based effective sampling techniques for enriching complex image samples into CNN training phases. This concludes the effective contribution of the proposed IGACM technique for detecting brain hemorrhages than the existing diagnosis models.

Artificial Intelligence in Pharmaceutical Field - A Critical Review

Artificial intelligence is an emerging sector in almost all fields. It is not confined only to a particular category and can be used in various fields like research, technology, and health. AI mainly concentrates on how computers analyze data and mimic the human thought process. As drug development involves high R & D costs and uncertainty in time consumption, artificial intelligence can serve as one of the promising solutions to overcome all these demerits. Due to the availability of enormous data, there are chances of missing out on some crucial details. For solving these issues, algorithms like machine learning, deep learning, and other expert systems are being used. On successful implementation of AI in the pharmaceutical field, the delays in drug development, and failure at the clinical and marketing level can be reduced. This review comprises information regarding the development of AI, its subfields, its overall implementation, and its application in the pharmaceutical sector and provides insights on challenges and limitations concerning AI.

Infrared Thermography as a Method of Verification in Raynaud's Phenomenon

Raynaud's phenomenon (RP) is characterized by the episodic whitening of the fingers upon exposure to cold. A recently described thermographic algorithm was proposed as a diagnostic replacement of the currently applied finger systolic pressure (FSP) test. The aim of the study was to evaluate the performance of the thermographic algorithm when applied in patients suspected of having RP. Forty-three patients were examined using thermographic imaging after local cooling of the hands in water of 10 °C for 1 min. The thermographic algorithm was applied to predict the probability of RP. The performance of the algorithm was evaluated with different cut-off levels. A new algorithm was proposed based on patients from the target population. The performance of the tested algorithm was noninferior to the FSP test, when a cut-off level of 0.05 was applied, yielding a sensitivity and specificity of 69% and 58%, respectively. The accuracy was 66%. The FSP test had a sensitivity and specificity of 77% and 37%, respectively, and the accuracy was 59%. The thermographic method proved useful for detecting RP and was able to replace the FSP test as a diagnostic test. The alternative algorithm revealed that other thermographic variables were more predictive of the target population, but this should be verified in future patients.

Characteristic Features of Infrared Thermographic Imaging in Primary Raynaud's Phenomenon

Raynaud’s phenomenon (RP) is characterized by the episodic whitening of the fingers upon exposure to cold. Verification of the condition is crucial in vibration-exposed patients. The current verification method is outdated, but thermographic imaging seems promising as a diagnostic replacement. By investigating patients diagnosed with RP, the study aimed at developing a simple thermographic procedure that could be applied to future patients where verification of the diagnosis is needed. Twenty-two patients with primary RP and 58 healthy controls were examined using thermographic imaging after local cooling of the hands for 1 min in water of 10°C. A logistic regression model was fitted with the temperature curve characteristics to convey a predicted probability of having RP. The characteristics time to end temperature and baseline temperature were the most appropriate predictors of RP among those examined (p = 0.004 and p = 0.04, respectively). The area under the curve was 0.91. The cut-off level 0.46 yielded a sensitivity and specificity of 82% and 86%, respectively. The positive and negative predictive values were 69% and 93%, respectively. This newly developed thermographic method was able to distinguish between patients with RP and healthy controls and was easy to operate. Thus, the method showed great promise as a method for verification of RP in future patients. Trial registration: ClinicalTrials.gov NCT03094910.