Enhancing Accuracy in Breast Density Assessment Using Deep Learning: A Multicentric, Multi-Reader Study.

The evaluation of mammographic breast density, a critical indicator of breast cancer risk, is traditionally performed by radiologists via visual inspection of mammography images, utilizing the Breast Imaging-Reporting and Data System (BI-RADS) breast density categories. However, this method is subject to substantial interobserver variability, leading to inconsistencies and potential inaccuracies in density assessment and subsequent risk estimations. To address this, we present a deep learning-based automatic detection algorithm (DLAD) designed for the automated evaluation of breast density. Our multicentric, multi-reader study leverages a diverse dataset of 122 full-field digital mammography studies (488 images in CC and MLO projections) sourced from three institutions. We invited two experienced radiologists to conduct a retrospective analysis, establishing a ground truth for 72 mammography studies (BI-RADS class A: 18, BI-RADS class B: 43, BI-RADS class C: 7, BI-RADS class D: 4). The efficacy of the DLAD was then compared to the performance of five independent radiologists with varying levels of experience. The DLAD showed robust performance, achieving an accuracy of 0.819 (95% CI: 0.736-0.903), along with an F1 score of 0.798 (0.594-0.905), precision of 0.806 (0.596-0.896), recall of 0.830 (0.650-0.946), and a Cohen's Kappa (κ) of 0.708 (0.562-0.841). The algorithm achieved robust performance that matches and in four cases exceeds that of individual radiologists. The statistical analysis did not reveal a significant difference in accuracy between DLAD and the radiologists, underscoring the model's competitive diagnostic alignment with professional radiologist assessments. These results demonstrate that the deep learning-based automatic detection algorithm can enhance the accuracy and consistency of breast density assessments, offering a reliable tool for improving breast cancer screening outcomes.

Leveraging Deep Learning Decision-Support System in Specialized Oncology Center: A Multi-Reader Retrospective Study on Detection of Pulmonary Lesions in Chest X-ray Images.

Chest X-ray (CXR) is considered to be the most widely used modality for detecting and monitoring various thoracic findings, including lung carcinoma and other pulmonary lesions. However, X-ray imaging shows particular limitations when detecting primary and secondary tumors and is prone to reading errors due to limited resolution and disagreement between radiologists. To address these issues, we developed a deep-learning-based automatic detection algorithm (DLAD) to automatically detect and localize suspicious lesions on CXRs. Five radiologists were invited to retrospectively evaluate 300 CXR images from a specialized oncology center, and the performance of individual radiologists was subsequently compared with that of DLAD. The proposed DLAD achieved significantly higher sensitivity (0.910 (0.854-0.966)) than that of all assessed radiologists (RAD 10.290 (0.201-0.379), p < 0.001, RAD 20.450 (0.352-0.548), p < 0.001, RAD 30.670 (0.578-0.762), p < 0.001, RAD 40.810 (0.733-0.887), p = 0.025, RAD 50.700 (0.610-0.790), p < 0.001). The DLAD specificity (0.775 (0.717-0.833)) was significantly lower than for all assessed radiologists (RAD 11.000 (0.984-1.000), p < 0.001, RAD 20.970 (0.946-1.000), p < 0.001, RAD 30.980 (0.961-1.000), p < 0.001, RAD 40.975 (0.953-0.997), p < 0.001, RAD 50.995 (0.985-1.000), p < 0.001). The study results demonstrate that the proposed DLAD could be utilized as a decision-support system to reduce radiologists' false negative rate.

Prospective evaluation of contrast-enhanced ultrasound of breast BI-RADS 3-5 lesions.

BACKGROUND: To determine the benefit of contrast-enhanced ultrasound (CEUS) in the assessment of breast lesions. METHODS: A standardized contrast-enhanced ultrasound was performed in 230 breast lesions classified as BI-RADS category 3 to 5. All lesions were subjected to qualitative and quantitative analysis. MVI (MicroVascular Imaging) technique was used to derive qualitative analysis parameters; blood perfusion of the lesions was assessed (perfusion homogeneity, type of vascularization, enhancement degree). Quantitative analysis was conducted to estimate perfusion changes in the lesions within drawn regions of interest (ROI); parameters TTP (time to peak), PI (peak intensity), WIS (wash in slope), AUC (area under curve) were obtained from time intensity (TI) curves. Acquired data were statistically analyzed to assess the ability of each parameter to differentiate between malignant and benign lesions. The combination of parameters was also evaluated for the possibility of increasing the overall diagnostic accuracy. Biological nature of the lesions was verified by a pathologist. Benign lesions without histopathological verification (BI-RADS 3) were followed up for at least 24 months. RESULTS: Out of 230 lesions, 146 (64%) were benign, 67 (29%) were malignant, 17 (7%) lesions were eliminated. Malignant tumors showed statistically significantly lower TTP parameters (sensitivity 77.6%, specificity 52.7%) and higher WIS values (sensitivity 74.6%, specificity 66.4%) than benign tumors. Enhancement degree also proved to be statistically well discriminating as 55.2% of malignant lesions had a rich vascularity (sensitivity 89.6% and specificity 48.6%). The combination of quantitative analysis parameters (TTP, WIS) with enhancement degree did not result in higher accuracy in distinguishing between malignant and benign breast lesions. CONCLUSIONS: We have demonstrated that contrast-enhanced breast ultrasound has the potential to distinguish between malignant and benign lesions. In particular, this method could help to differentiate lesions BI-RADS category 3 and 4 and thus reduce the number of core-cut biopsies performed in benign lesions. Qualitative analysis, despite its subjective element, appeared to be more beneficial. A combination of quantitative and qualitative analysis did not increase the predictive capability of CEUS.

In silico study of pseudoprogression in glioblastoma: collaboration of radiologists and radiation oncologists in the estimation of extent of high dose RT region.

BACKGROUND AND AIM: Oncologists play a vital role in the interpretation of radiographic results in glioblastoma patients. Molecular pathology and information on radiation treatment protocols among others are all important for accurate interpretation of radiology images. One important issue that may arise in interpreting such images is the phenomenon of tumor "pseudoprogression"; oncologists need to be able to distinguish this effect from true disease progression.Exact knowledge about the location of high-dose radiotherapy region is needed for valid determination of pseudoprogression according to RANO (Response Assessment in Neuro-Oncology) criteria in neurooncology. The aim of the present study was to evaluate the radiologists' understanding of a radiotherapy high-dose region in routine clinical practice since radiation oncologists do not always report 3-dimensional isodoses when ordering follow up imaging. METHODS: Eight glioblastoma patients who underwent postresection radiotherapy were included in this study. Four radiologists worked with their pre-radiotherapy planning MR, however, they were blinded to RT target volumes which were defined by radiation oncologists according to current guidelines. The aim was to draw target volume for high dose RT fields (that is the region, where they would consider that there may be a pseudoprogression in future MRI scans). Many different indices describing structure differences were analyzed in comparison with original per-protocol RT target volumes. RESULTS: The median volume for RT high dose field was 277 ccm (range 218 to 401 ccm) as defined per protocol by radiation oncologist and 87 ccm (range 32-338) as defined by radiologists (median difference of paired difference 31%, range 15-112%). The Median Dice index of similarity was 0.46 (range 0.14 - 0.78), the median Hausdorff distance 25 mm. CONCLUSION: Continuing effort to improve education on specific procedures in RT and in radiology as well as automatic tools for exporting RT targets is needed in order to increase specificity and sensitivity in response evaluation.