Assessing the malignancy of suspicious breast microcalcifications: the role of contrast enhanced mammography.

PURPOSE: To assess the role of contrast-enhanced mammography (CEM) in predicting the malignancy of breast calcifications. MATERIAL AND METHODS: We retrospectively evaluated patients with suspicious calcifications (BIRADS 4) who underwent CEM and stereotactic vacuum-assisted biopsy (VAB) at our institution. We assessed the sensitivity (SE), specificity (SP), positive predictive value (PPV) and negative predictive value (NPV) of CEM in predicting malignancy of microcalcifications with a 95% confidence interval; we performed an overall analysis and a subgroup analysis stratified into group A-low risk (BIRADS 4a) and group B-medium/high risk (BIRADS 4b-4c). We then evaluated the correlation between enhancement and tumour proliferation index (Ki-67) for all malignant lesions. RESULTS: Data from 182 patients with 184 lesions were collected. Overall the SE of CEM in predicting the malignancy of microcalcifications was 0.70, SP was 0.85, the PPV was 0.82, the NPV was 0.76 and AUC was 0.78. SE in group A was 0.89, SP was 0.89, PPV was 0.57, NPV was 0.98 and AUC was 0.75. SE in group B was 0.68, SP was 0.80, PPV was 0.87, NPV was 0.57 and AUC was 0.75. Among malignant microcalcifications that showed enhancement (N = 52), 61.5% had Ki-67 ≥ 20% and 38.5% had low Ki-67 values. Among the lesions that did not show enhancement (N = 22), 90.9% had Ki-67 < 20% and 9.1% showed high Ki-67 values 20%. CONCLUSIONS: The absence of enhancement can be used as an indicative parameter for the absence of disease in cases of low-suspicious microcalcifications, but not in intermediate-high suspicious ones for which biopsy remains mandatory and can be used to distinguish indolent lesions from more aggressive neoplasms, with consequent reduction of overdiagnosis and overtreatment.

How do large language models answer breast cancer quiz questions? A comparative study of GPT-3.5, GPT-4 and Google Gemini.

Applications of large language models (LLMs) in the healthcare field have shown promising results in processing and summarizing multidisciplinary information. This study evaluated the ability of three publicly available LLMs (GPT-3.5, GPT-4, and Google Gemini-then called Bard) to answer 60 multiple-choice questions (29 sourced from public databases, 31 newly formulated by experienced breast radiologists) about different aspects of breast cancer care: treatment and prognosis, diagnostic and interventional techniques, imaging interpretation, and pathology. Overall, the rate of correct answers significantly differed among LLMs (p = 0.010): the best performance was achieved by GPT-4 (95%, 57/60) followed by GPT-3.5 (90%, 54/60) and Google Gemini (80%, 48/60). Across all LLMs, no significant differences were observed in the rates of correct replies to questions sourced from public databases and newly formulated ones (p ≥ 0.593). These results highlight the potential benefits of LLMs in breast cancer care, which will need to be further refined through in-context training.

Association between mpMRI detected tumor apparent diffusion coefficient and 5-year biochemical recurrence risk after radical prostatectomy.

PURPOSE: To assess the ability of tumor apparent diffusion coefficient (ADC) values obtained from multiparametric magnetic resonance imaging (mpMRI) to predict the risk of 5-year biochemical recurrence (BCR) after radical prostatectomy (RP). MATERIALS AND METHODS: This retrospective analysis included 1207 peripheral and 232 non-peripheral zone prostate cancer (PCa) patients who underwent mpMRI before RP (2012-2015), with the outcome of interest being 5-year BCR. ADC was evaluated as a continuous variable and as categories: low (< 850 µm2/s), intermediate (850-1100 µm2/s), and high (> 1100 µm2/s). Kaplan-Meier curves with log-rank testing of BCR-free survival, multivariable Cox proportional hazard regression models were formed to estimate the risk of BCR. RESULTS: Among the 1439 males with median age 63 (± 7) years, the median follow-up was 59 months, and 306 (25%) patients experienced BCR. Peripheral zone PCa patients with BCR had lower tumor ADC values than those without BCR (874 versus 1025 µm2/s, p < 0.001). Five-year BCR-free survival rates were 52.3%, 74.4%, and 87% for patients in the low, intermediate, and high ADC value categories, respectively (p < 0.0001). Lower ADC was associated with BCR, both as continuously coded variable (HR: 5.35; p < 0.001) and as ADC categories (intermediate versus high ADC-HR: 1.56, p = 0.017; low vs. high ADC-HR; 2.36, p < 0.001). In the non-peripheral zone PCa patients, no association between ADC and BCR was observed. CONCLUSION: Tumor ADC values and categories were found to be predictive of the 5-year BCR risk after RP in patients with peripheral zone PCa and may serve as a prognostic biomarker.

Exploring the Potential of Artificial Intelligence in Breast Ultrasound.

Breast ultrasound has emerged as a valuable imaging modality in the detection and characterization of breast lesions, particularly in women with dense breast tissue or contraindications for mammography. Within this framework, artificial intelligence (AI) has garnered significant attention for its potential to improve diagnostic accuracy in breast ultrasound and revolutionize the workflow. This review article aims to comprehensively explore the current state of research and development in harnessing AI's capabilities for breast ultrasound. We delve into various AI techniques, including machine learning, deep learning, as well as their applications in automating lesion detection, segmentation, and classification tasks. Furthermore, the review addresses the challenges and hurdles faced in implementing AI systems in breast ultrasound diagnostics, such as data privacy, interpretability, and regulatory approval. Ethical considerations pertaining to the integration of AI into clinical practice are also discussed, emphasizing the importance of maintaining a patient-centered approach. The integration of AI into breast ultrasound holds great promise for improving diagnostic accuracy, enhancing efficiency, and ultimately advancing patient's care. By examining the current state of research and identifying future opportunities, this review aims to contribute to the understanding and utilization of AI in breast ultrasound and encourage further interdisciplinary collaboration to maximize its potential in clinical practice.