Label-Efficient Self-Supervised Federated Learning for Tackling Data Heterogeneity in Medical Imaging

Yan, Rui; Qu, Liangqiong; Wei, Qingyue; Huang, Shih-Cheng; Shen, Liyue; Rubin, Daniel L.; Liu, Xing; Zhou, Yuyin

doi:10.1109/tmi.2022.3233574

Cited by 57 publications

(15 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A variety of emerging methods, including federated learning (43), synthetic data augmentation and semi-supervised approaches (30), active learning (13), and AI-assisted annotation or contour editing (7,16,17) are being explored for overcoming the scarcity of labeled data in cross-sectional imaging (6). In the meantime, investigators are beginning to use collaborative human-in-the-loop strategies with supervision and editing of automated labels to accelerate voxelwise annotation (19)(20)(21).…”

Section: Discussionmentioning

confidence: 99%

Accelerating voxelwise annotation of cross-sectional imaging through AI collaborative labeling with quality assurance and bias mitigation

et al. 2023

View full text Add to dashboard Cite

Backgroundprecision-medicine quantitative tools for cross-sectional imaging require painstaking labeling of targets that vary considerably in volume, prohibiting scaling of data annotation efforts and supervised training to large datasets for robust and generalizable clinical performance. A straight-forward time-saving strategy involves manual editing of AI-generated labels, which we call AI-collaborative labeling (AICL). Factors affecting the efficacy and utility of such an approach are unknown. Reduction in time effort is not well documented. Further, edited AI labels may be prone to automation bias.PurposeIn this pilot, using a cohort of CTs with intracavitary hemorrhage, we evaluate both time savings and AICL label quality and propose criteria that must be met for using AICL annotations as a high-throughput, high-quality ground truth.Methods57 CT scans of patients with traumatic intracavitary hemorrhage were included. No participant recruited for this study had previously interpreted the scans. nnU-net models trained on small existing datasets for each feature (hemothorax/hemoperitoneum/pelvic hematoma; n = 77–253) were used in inference. Two common scenarios served as baseline comparison- de novo expert manual labeling, and expert edits of trained staff labels. Parameters included time effort and image quality graded by a blinded independent expert using a 9-point scale. The observer also attempted to discriminate AICL and expert labels in a random subset (n = 18). Data were compared with ANOVA and post-hoc paired signed rank tests with Bonferroni correction.ResultsAICL reduced time effort 2.8-fold compared to staff label editing, and 8.7-fold compared to expert labeling (corrected p < 0.0006). Mean Likert grades for AICL (8.4, SD:0.6) were significantly higher than for expert labels (7.8, SD:0.9) and edited staff labels (7.7, SD:0.8) (corrected p < 0.0006). The independent observer failed to correctly discriminate AI and human labels.ConclusionFor our use case and annotators, AICL facilitates rapid large-scale curation of high-quality ground truth. The proposed quality control regime can be employed by other investigators prior to embarking on AICL for segmentation tasks in large datasets.

show abstract

Section: Discussionmentioning

confidence: 99%

Accelerating voxelwise annotation of cross-sectional imaging through AI collaborative labeling with quality assurance and bias mitigation

et al. 2023

View full text Add to dashboard Cite

show abstract

“…Some studies have realized the development of domain adaptation and domain generalization algorithms based on the FL framework and applied them to multi-center datasets under patient privacy protection. 128,[146][147][148] The results of these studies indicate that designing novel federated domain adaptation or generalization methods is a promising research direction for medical image analysis. They can maintain the accuracy of the predictive models while protecting patient privacy.…”

Section: Patient Privacy Protectionmentioning

confidence: 99%

Improving prediction of treatment response and prognosis in colorectal cancer with AI-based medical image analysis

Liu,

Zhang,

Shao

et al. 2024

TIME

View full text Add to dashboard Cite

<p>The heterogeneous response and prognosis of patients with colorectal cancer (CRC) to standard treatment regimens remains a challenge for clinical management. Individually weak prognostic markers, defined by gene mutations and protein expression, are difficult to apply in routine clinical practice because of their high acquisition cost and mediocre prediction accuracy. Visual evaluation of medical images, including radiology and digital pathology images, is an important part of CRC management. With the rapid development of artificial intelligence (AI), high-dimensional imaging features other than visual information are increasingly being used to develop imaging markers. At different stages of treatment, accurate predictions of treatment response and prognosis may help in selecting patients and tailoring their treatment. Here, we review the current state of AI applied to the medical imaging of CRC and describe its recent progress in short-term response and long-term survival prediction. In addition, we illustrate how these AI-based approaches may affect clinical decision-making. Although few approaches have been applied in routine clinical practice, their results are promising. Finally, we discuss the challenges in applying AI in clinical practice and possible future solutions from three perspectives: model interpretability, model generalizability, and patient privacy protection. This comprehensive assessment underscores the transformative potential of AI in CRC management and emphasizes the need for further exploration and integration into routine clinical workflows.</p>

show abstract

“…It is frequently up-dated to provide information about the skin via any desktop or mobile web browser. Moreover, high-definition, non-watermarked images are available for purchase [108,117,119].…”

Section: Private Datasetsmentioning

confidence: 99%

Federated and Transfer Learning Methods for the Classification of Melanoma and Nonmelanoma Skin Cancers: A Prospective Study

Riaz,

Naeem,

Malik

et al. 2023

Sensors

View full text Add to dashboard Cite

Skin cancer is considered a dangerous type of cancer with a high global mortality rate. Manual skin cancer diagnosis is a challenging and time-consuming method due to the complexity of the disease. Recently, deep learning and transfer learning have been the most effective methods for diagnosing this deadly cancer. To aid dermatologists and other healthcare professionals in classifying images into melanoma and nonmelanoma cancer and enabling the treatment of patients at an early stage, this systematic literature review (SLR) presents various federated learning (FL) and transfer learning (TL) techniques that have been widely applied. This study explores the FL and TL classifiers by evaluating them in terms of the performance metrics reported in research studies, which include true positive rate (TPR), true negative rate (TNR), area under the curve (AUC), and accuracy (ACC). This study was assembled and systemized by reviewing well-reputed studies published in eminent fora between January 2018 and July 2023. The existing literature was compiled through a systematic search of seven well-reputed databases. A total of 86 articles were included in this SLR. This SLR contains the most recent research on FL and TL algorithms for classifying malignant skin cancer. In addition, a taxonomy is presented that summarizes the many malignant and non-malignant cancer classes. The results of this SLR highlight the limitations and challenges of recent research. Consequently, the future direction of work and opportunities for interested researchers are established that help them in the automated classification of melanoma and nonmelanoma skin cancers.

show abstract

Label-Efficient Self-Supervised Federated Learning for Tackling Data Heterogeneity in Medical Imaging

Cited by 57 publications

References 29 publications

Accelerating voxelwise annotation of cross-sectional imaging through AI collaborative labeling with quality assurance and bias mitigation

Accelerating voxelwise annotation of cross-sectional imaging through AI collaborative labeling with quality assurance and bias mitigation

Improving prediction of treatment response and prognosis in colorectal cancer with AI-based medical image analysis

Federated and Transfer Learning Methods for the Classification of Melanoma and Nonmelanoma Skin Cancers: A Prospective Study

Contact Info

Product

Resources

About