RESUMEN
RATIONALE AND OBJECTIVES: The aim of this study was to develop a deep learning radiomics nomogram (DLRN) based on B-mode ultrasound (BMUS) and color doppler flow imaging (CDFI) images for preoperative assessment of lymphovascular invasion (LVI) status in invasive breast cancer (IBC). MATERIALS AND METHODS: In this multicenter, retrospective study, 832 pathologically confirmed IBC patients were recruited from eight hospitals. The samples were divided into training, internal test, and external test sets. Deep learning and handcrafted radiomics features reflecting tumor phenotypes on BMUS and CDFI images were extracted. The BMUS score and CDFI score were calculated after radiomics feature selection. Subsequently, a DLRN was developed based on the scores and independent clinic-ultrasonic risk variables. The performance of the DLRN was evaluated for calibration, discrimination, and clinical usefulness. RESULTS: The DLRN predicted the LVI with accuracy, achieving an area under the receiver operating characteristic curve of 0.93 (95% CI 0.90-0.95), 0.91 (95% CI 0.87-0.95), and 0.91 (95% CI 0.86-0.94) in the training, internal test, and external test sets, respectively, with good calibration. The DLRN demonstrated superior performance compared to the clinical model and single scores across all three sets (p < 0.05). Decision curve analysis and clinical impact curve confirmed the clinical utility of the model. Furthermore, significant enhancements in net reclassification improvement (NRI) and integrated discrimination improvement (IDI) indicated that the two scores could serve as highly valuable biomarkers for assessing LVI. CONCLUSION: The DLRN exhibited strong predictive value for LVI in IBC, providing valuable information for individualized treatment decisions.
RESUMEN
The aim of this retrospective study was to evaluate the efficacy of contrast-enhanced ultrasound (CEUS) washout rate in predicting hepatocellular carcinoma (HCC) differentiation. Two hundred seventy-one patients underwent liver resection for HCC between April 2008 and December 2012 after being examined by CEUS using the contrast agent SonoVue with a low mechanical index (<0.1) in a routine procedure. Contrast agent washout rates obtained from video images were divided into four categories from slow to fast: WR1 = no washout in all phases (slowest); WR2 = washout after 120 s from contrast injection (late-phase washout); WR3 = washout between 41 and 120 s from contrast injection (portal venous washout); WR4 = washout before 40 s from contrast injection (fastest washout rate). HCC nodules were graded as well, moderately and poorly differentiated. Spearman rank correlation and χ(2)-tests were used to assess group relationships and differences. Receiver operating characteristic curve analysis was used to determine the diagnostic predictive value of CEUS. Among the 271 patients, 18 (6.6%) had well differentiated, 150 (55.4%) had moderately differentiated and 103 (38.0%) had poorly differentiated HCC. Statistical tests indicated that washout rate was significantly correlated with tumor differentiation (p < 0.05), and the poorly differentiated HCCs had earlier washout. At the cutoff point of WR4, CEUS based on washout rate performed poorly in distinguishing poorly differentiated from moderately and well-differentiated HCCs, with a sensitivity, specificity and accuracy (area under the curve) of 24%, 97% and 0.68, respectively. However, at the cutoff point of WR2, the sensitivity, specificity and accuracy of CEUS in differentiating well-differentiated HCC from other HCCs were significantly better: 98%, 78% and 0.96, respectively. Thus, CEUS washout rate may have a role in identifying patients with well-differentiated HCC.
