Enhancing the fairness of AI prediction models by Quasi-Pareto improvement among heterogeneous thyroid nodule population.

Artificial Intelligence (AI) models for medical diagnosis often face challenges of generalizability and fairness. We highlighted the algorithmic unfairness in a large thyroid ultrasound dataset with significant diagnostic performance disparities across subgroups linked causally to sample size imbalances. To address this, we introduced the Quasi-Pareto Improvement (QPI) approach and a deep learning implementation (QP-Net) combining multi-task learning and domain adaptation to improve model performance among disadvantaged subgroups without compromising overall population performance. On the thyroid ultrasound dataset, our method significantly mitigated the area under curve (AUC) disparity for three less-prevalent subgroups by 0.213, 0.112, and 0.173 while maintaining the AUC for dominant subgroups; we also further confirmed the generalizability of our approach on two public datasets: the ISIC2019 skin disease dataset and the CheXpert chest radiograph dataset. Here we show the QPI approach to be widely applicable in promoting AI for equitable healthcare outcomes.

tRFtarget 2.0: expanding the targetome landscape of transfer RNA-derived fragments.

tRFtarget 1.0 (http://trftarget.net/) is a platform consolidating both computationally predicted and experimentally validated binding sites between transfer RNA-derived fragments (tRFs) and target genes (or transcripts) across multiple organisms. Here, we introduce a newly released version of tRFtarget 2.0, in which we integrated 6 additional tRF sources, resulting in a comprehensive collection of 2614 high-quality tRF sequences spanning across 9 species, including 1944 Homo sapiens tRFs and one newly incorporated species Rattus norvegicus. We also expanded target genes by including ribosomal RNAs, long non-coding RNAs, and coding genes >50 kb in length. The predicted binding sites have surged up to approximately 6 billion, a 20.5-fold increase than that in tRFtarget 1.0. The manually curated publications relevant to tRF targets have increased to 400 and the gene-level experimental evidence has risen to 232. tRFtarget 2.0 introduces several new features, including a web-based tool that identifies potential binding sites of tRFs in user's own datasets, integration of standardized tRF IDs, and inclusion of external links to contents within the database. Additionally, we enhanced website framework and user interface. With these improvements, tRFtarget 2.0 is more user-friendly, providing researchers a streamlined and comprehensive platform to accelerate their research progress.

Automating General Movements Assessment with quantitative deep learning to facilitate early screening of cerebral palsy.

The Prechtl General Movements Assessment (GMA) is increasingly recognized for its role in evaluating the integrity of the developing nervous system and predicting motor dysfunctions, particularly in conditions such as cerebral palsy (CP). However, the necessity for highly trained professionals has hindered the adoption of GMA as an early screening tool in some countries. In this study, we propose a deep learning-based motor assessment model (MAM) that combines infant videos and basic characteristics, with the aim of automating GMA at the fidgety movements (FMs) stage. MAM demonstrates strong performance, achieving an Area Under the Curve (AUC) of 0.967 during external validation. Importantly, it adheres closely to the principles of GMA and exhibits robust interpretability, as it can accurately identify FMs within videos, showing substantial agreement with expert assessments. Leveraging the predicted FMs frequency, a quantitative GMA method is introduced, which achieves an AUC of 0.956 and enhances the diagnostic accuracy of GMA beginners by 11.0%. The development of MAM holds the potential to significantly streamline early CP screening and revolutionize the field of video-based quantitative medical diagnostics.

Human understandable thyroid ultrasound imaging AI report system - A bridge between AI and clinicians.

Artificial intelligence (AI) enables accurate diagnosis of thyroid cancer; however, the lack of explanation limits its application. In this study, we collected 10,021 ultrasound images from 8,079 patients across four independent institutions to develop and validate a human understandable AI report system named TiNet for thyroid cancer prediction. TiNet can extract thyroid nodule features such as texture, margin, echogenicity, shape, and location using a deep learning method conforming to the clinical diagnosis standard. Moreover, it offers excellent prediction performance (AUC 0.88) and provides quantitative explanations for the predictions. We conducted a reverse cognitive test in which clinicians matched the correct ultrasound images according to TiNet and clinical reports. The results indicated that TiNet reports (87.1% accuracy) were significantly easier to understand than clinical reports (81.6% accuracy; p < 0.001). TiNet can serve as a bridge between AI-based diagnosis and clinicians, enhancing human-AI cooperative medical decision-making.

Automated diagnosis and management of follicular thyroid nodules based on the devised small-dataset interpretable foreground optimization network deep learning: a multicenter diagnostic study.

BACKGROUND: Currently, follicular thyroid carcinoma (FTC) has a relatively low incidence with a lack of effective preoperative diagnostic means. To reduce the need for invasive diagnostic procedures and to address information deficiencies inherent in a small dataset, we utilized interpretable foreground optimization network deep learning to develop a reliable preoperative FTC detection system. METHODS: In this study, a deep learning model (FThyNet) was established using preoperative ultrasound images. Data on patients in the training and internal validation cohort ( n =432) were obtained from Ruijin Hospital, China. Data on patients in the external validation cohort ( n =71) were obtained from four other clinical centers. We evaluated the predictive performance of FThyNet and its ability to generalize across multiple external centers and compared the results yielded with assessments from physicians directly predicting FTC outcomes. In addition, the influence of texture information around the nodule edge on the prediction results was evaluated. RESULTS: FThyNet had a consistently high accuracy in predicting FTC with an area under the receiver operating characteristic curve (AUC) of 89.0% [95% CI 87.0-90.9]. Particularly, the AUC for grossly invasive FTC reached 90.3%, which was significantly higher than that of the radiologists (56.1% [95% CI 51.8-60.3]). The parametric visualization study found that those nodules with blurred edges and relatively distorted surrounding textures were more likely to have FTC. Furthermore, edge texture information played an important role in FTC prediction with an AUC of 68.3% [95% CI 61.5-75.5], and highly invasive malignancies had the highest texture complexity. CONCLUSION: FThyNet could effectively predict FTC, provide explanations consistent with pathological knowledge, and improve clinical understanding of the disease.

A meta-learning approach to improving radiation response prediction in cancers.

PURPOSE: Predicting the efficacy of radiotherapy in individual patients has drawn widespread attention, but the limited sample size remains a bottleneck for utilizing high-dimensional multi-omics data to guide personalized radiotherapy. We hypothesize the recently developed meta-learning framework could address this limitation. METHODS AND MATERIALS: By combining gene expression, DNA methylation, and clinical data of 806 patients who had received radiotherapy from The Cancer Genome Atlas (TCGA), we applied the Model-Agnostic Meta-Learning (MAML) framework to tasks consisting of pan-cancer data, to obtain the best initial parameters of a neural network for a specific cancer with smaller number of samples. The performance of meta-learning framework was compared with four traditional machine learning methods based on two training schemes, and tested on Cancer Cell Line Encyclopedia (CCLE) and Chinese Glioma Genome Atlas (CGGA) datasets. Moreover, biological significance of the models was investigated by survival analysis and feature interpretation. RESULTS: The mean AUC (Area under the ROC Curve) [95% confidence interval] of our models across nine cancer types was 0.702 [0.691-0.713], which improved by 0.166 on average over other the four machine learning methods on two training schemes. Our models performed significantly better (p < 0.05) in seven cancer types and performed comparable to the other predictors in the rest of two cancer types. The more pan-cancer samples were used to transfer meta-knowledge, the greater the performance improved (p < 0.05). The predicted response scores that our models generated were negatively correlated with cell radiosensitivity index in four cancer types (p < 0.05), while not statistically significant in the other three cancer types. Moreover, the predicted response scores were shown to be prognostic factors in seven cancer types and eight potential radiosensitivity-related genes were identified. CONCLUSIONS: For the first time, we established the meta-learning approach to improving individual radiation response prediction by transferring common knowledge from pan-cancer data with MAML framework. The results demonstrated the superiority, generalizability, and biological significance of our approach.

Texture Synthesis Based Thyroid Nodule Detection From Medical Ultrasound Images: Interpreting and Suppressing the Adversarial Effect of In-place Manual Annotation.

Deep learning method have been offering promising solutions for medical image processing, but failing to understand what features in the input image are captured and whether certain artifacts are mistakenly included in the model, thus create crucial problems in generalizability of the model. We targeted a common issue of this kind caused by manual annotations appeared in medical image. These annotations are usually made by the doctors at the spot of medical interest and have adversarial effect on many computer vision AI tasks. We developed an inpainting algorithm to remove the annotations and recover the original images. Besides we applied variational information bottleneck method in order to filter out the unwanted features and enhance the robustness of the model. Our impaiting algorithm is extensively tested in object detection in thyroid ultrasound image data. The mAP (mean average precision, with IoU = 0.3) is 27% without the annotation removal. The mAP is 83% if manually removed the annotations using Photoshop and is enhanced to 90% using our inpainting algorithm. Our work can be utilized in future development and evaluation of artificial intelligence models based on medical images with defects.