The volume of medical images stored in hospitals is rapidly increasing; however, the utilization of these accumulated medical images remains limited. Existing content-based medical image retrieval (CBMIR) systems typically require example images, leading to practical limitations, such as the lack of customizable, fine-grained image retrieval, the inability to search without example images, and difficulty in retrieving rare cases. In this paper, we introduce a sketch-based medical image retrieval (SBMIR) system that enables users to find images of interest without the need for example images. The key concept is feature decomposition of medical images, which allows the entire feature of a medical image to be decomposed into and reconstructed from normal and abnormal features. Building on this concept, our SBMIR system provides an easy-to-use two-step graphical user interface: users first select a template image to specify a normal feature and then draw a semantic sketch of the disease on the template image to represent an abnormal feature. The system integrates both types of input to construct a query vector and retrieves reference images. For evaluation, ten healthcare professionals participated in a user test using two datasets. Consequently, our SBMIR system enabled users to overcome previous challenges, including image retrieval based on fine-grained image characteristics, image retrieval without example images, and image retrieval for rare cases. Our SBMIR system provides on-demand, customizable medical image retrieval, thereby expanding the utility of medical image databases.

Objective physical function assessment is crucial for determining patient eligibility for treatment and adjusting the treatment intensity. Existing assessments, such as performance status, are not well standardized, despite their frequent use in daily clinical practice. This paper explored how artificial intelligence (AI) could predict physical function scores from various patient data sources and reviewed methods to measure objective physical function using this technology. This review included relevant articles published in English that were retrieved from PubMed. These studies utilized AI technology to predict physical function indices from patient data extracted from videos, sensors, or electronic health records, thereby eliminating manual measurements. Studies that used AI technology solely to automate traditional evaluations were excluded. These technologies are recommended for future clinical systems that perform repeated objective physical function assessments in all patients without requiring extra time, personnel, or resources. This enables the detection of minimal changes in a patient’s condition, enabling early intervention and enhanced outcomes.

Fetal cardiac ultrasound screening is vital for the prenatal diagnosis of congenital heart disease (CHD). However, understanding complex cardiovascular structures and continuity from two-dimensional (2D) ultrasound images remains challenging. Therefore, this study developed and evaluated an artificial intelligence (AI)-based three-dimensional (3D) visualization approach that reconstructs fetal cardiovascular structures from 2D ultrasound sweep video. To this end, information from the adjacent frames was used. Instead of interpreting each ultrasound frame independently, this method incorporates neighboring frames, enabling the continuous visualization of the right- and left-sided functional circulatory pathways. Our method achieved a mean Dice coefficient of 0.785 on an independent test dataset, demonstrating that the adjacent frame information can provide stable 3D segmentation. Diffusion-based anatomical plausibility analysis trained exclusively on normal anatomies showed exploratory discrimination between normal fetuses and fetuses with CHD (AUC, 0.861; 95% CI, 0.695–0.980). In a clinical comparison study involving 12 obstetricians, the introduction of this AI-driven 3D cardiovascular visualization slightly improved the screening accuracy from 0.779 to 0.808 and significantly improved the confidence-weighted accuracy score (p = 0.002). Our approach can potentially support examiners and facilitate fetal cardiac ultrasound screenings.

Background: PREDICT is a prognostic tool developed in the United Kingdom to estimate postoperative overall survival (OS) and the additional benefits of adjuvant therapies in patients with breast cancer. It has been validated in various international cohorts and continuously updated with the inclusion of new variables and model retraining. However, their efficacy in the Japanese population remains unclear. We aimed to evaluate the generalizability of PREDICT versions 2.2 (v2.2) and 3.0 (v3.0) using data from the National Cancer Center Hospital in Japan, a high-volume cancer center. Methods: We analyzed a retrospective cohort (2006–2016) including 2,980 cases of postoperative breast cancer. We calculated survival predictions using both v2.2 and v3.0, and compared them with the Kaplan–Meier-estimated survival probabilities using a calibration plot. Additionally, we performed a time-dependent receiver operating characteristic (ROC) curve analysis for v2.2 and v3.0. Results: Both models tended to underestimate survival in our cohort, whereas v3.0 showed improved calibration compared to v2.2, for 5- and 10-year OS. Both v2.2 and v3.0 maintained good discriminative performance throughout 10 years, with values under the ROC curve generally above 0.80. Conclusions: Despite these differences, both versions demonstrated satisfactory performance, suggesting that they can be generalized for Japanese patients with postoperative breast cancer. Notably, v3.0, given its improved calibration, might be more suitable for supporting shared decision making. The model’s performance in predicting 5-year OS supports its generalizability, whereas 10-year projections warrant caution due to limited follow-up. Overall, this study demonstrates that PREDICT is a valuable prognostic tool for Japanese patients with breast cancer.

The amount of medical images stored in hospitals is increasing faster than ever; however, utilizing the accumulated medical images has been limited. This is because existing content-based medical image retrieval (CBMIR) systems usually require example images to construct query vectors; nevertheless, example images cannot always be prepared. Besides, there can be images with rare characteristics that make it difficult to find similar example images, which we call isolated samples. Here, we introduce a novel sketch-based medical image retrieval (SBMIR) system that enables users to find images of interest without example images. The key idea lies in feature decomposition of medical images, whereby the entire feature of a medical image can be decomposed into and reconstructed from normal and abnormal features. By extending this idea, our SBMIR system provides an easy-to-use two-step graphical user interface: users first select a template image to specify a normal feature and then draw a semantic sketch of the disease on the template image to represent an abnormal feature. Subsequently, it integrates the two kinds of input to construct a query vector and retrieves reference images with the closest reference vectors. Using two datasets, ten healthcare professionals with various clinical backgrounds participated in the user test for evaluation. As a result, our SBMIR system enabled users to overcome previous challenges, including image retrieval based on fine-grained image characteristics, image retrieval without example images, and image retrieval for isolated samples. Our SBMIR system achieves flexible medical image retrieval on demand, thereby expanding the utility of medical image databases.

Assessing objective physical function in patients with cancer is crucial for evaluating their ability to tolerate invasive treatments. Current assessment methods, such as the timed up and go (TUG) test and the short physical performance battery, tend to require additional resources and time, limiting their practicality in routine clinical practice. To address these challenges, we developed a system to assess physical function based on movements observed during clinical consultations and aimed to explore relevant features from inertial measurement unit data collected during those movements. As for the flow of the research, we first collected inertial measurement unit data from 61 patients with cancer while they replicated a series of movements in a consultation room. We then conducted correlation analyses to identify keypoints of focus and developed machine learning models to predict the TUG test outcomes using the extracted features. Regarding results, pelvic velocity variability (PVV) was identified using Lasso regression. A linear regression model using PVV as the input variable achieved a mean absolute error of 1.322 s and a correlation of 0.713 with the measured TUG results during five-fold cross-validation. Higher PVV correlated with shorter TUG test results. These findings provide a foundation for the development of an artificial intelligence-based physical function assessment system that operates without the need for additional resources.