Transformer-based methods are recently popular in vision tasks because of their capability to model global dependencies alone. However, it limits the performance of networks due to the lack of modeling local context and global-local correlations of multi-scale features. In this paper, we present MISSFormer, a Medical Image Segmentation tranSFormer. MISSFormer is a hierarchical encoder-decoder network with two appealing designs: 1) a feed-forward network in transformer block of U-shaped encoder-decoder structure is redesigned, ReMix-FFN, which explore global dependencies and local context for better feature discrimination by re-integrating the local context and global dependencies; 2) a ReMixed Transformer Context Bridge is proposed to extract the correlations of global dependencies and local context in multi-scale features generated by our hierarchical transformer encoder. The MISSFormer shows a solid capacity to capture more discriminative dependencies and context in medical image segmentation. The experiments on multi-organ, cardiac segmentation and retinal vessel segmentation tasks demonstrate the superiority, effectiveness and robustness of our MISSFormer. Specifically, the experimental results of MISSFormer trained from scratch even outperform state-of-the-art methods pre-trained on ImageNet, and the core designs can be generalized to other visual segmentation tasks. The code has been released on Github: https://github.com/ZhifangDeng/MISSFormer.

Background and Objectives: To identify factors that influence the sample adequacy of solid thyroid nodules based on ultrasound-guided fine-needle aspiration (FNA) with subsequent liquid-based cytology. Materials and Methods: We retrospectively reviewed 855 patients who underwent ultrasound-guided FNA at our hospital between July 2019 and July 2020. The final analysis included 801 solid thyroid nodules in 801 patients. After reviewing the demographic data, ultrasonic features, and FNA technique-related factors, we defined 14 potential variables. For cytological results, the Bethesda categories II–VI were defined as adequate sample results. Univariate and multivariate analyses were performed to identify factors that influenced sample adequacy. Results: The adequate sample rate was 87.1%. The univariate analysis showed that four factors were related to adequate sampling in patients with thyroid FNA. These factors included age (p < 0.001), nodule orientation (p = 0.0232), calcification (p = 0.0034), and operator experience (p = 0.0286). After the multivariate analysis, five independent factors were identified to improve the diagnostic results of FNA for solid thyroid nodules: (1) the presence of Hashimoto’s thyroiditis (odds ratio (OR) = 1.810; 95% confidence interval (CI): 1.076–3.045; p = 0.0254), (2) a taller-than-wide orientation (OR = 2.038; 95% CI: 1.260–3.296; p = 0.0037), (3) the presence of calcification (OR = 1.767; 95% CI: 1.115–2.799; p = 0.0153), (4) four needle passes to obtain material (OR = 1.750; 95% CI: 1.094–2.799; p = 0.0196), and (5) an experienced operator (OR = 0.561; 95% CI: 0.319–0.987; p = 0.0451). Conclusions: A taller-than-wide orientation, the presence of calcification, and the presence of Hashimoto’s thyroiditis were found to affect the sample adequacy of ultrasound-guided FNA with liquid-based cytology. The sample adequacy could be improved when FNA is performed with four needle passes by experienced doctors.

Background: Subpleural focal organizing pneumonia (FOP) and primary lung malignancy (PLM) are usually confused. The aim of this study was to explore the value of ultrasound (US) and contrast-enhanced ultrasound (CEUS) in the differential diagnosis of FOP and PLM. Methods: A total of 23 patients (mean age: 64.57 ± 11.86 years) with FOP and 100 (mean age: 66.29 ± 11.05 years) with subpleural lesions diagnosed as PLM, confirmed by pathological diagnosis and clinical follow-up, were retrospectively enrolled. The largest lesion diameter, angle between the lesion border and thoracic wall, air bronchial sign, internal blood supply, blood supply form, and pleural effusion examined using conventional US were retrospectively analyzed. The indicators of CEUS included the arrival time of contrast agent in the lesion, lesion–lung arrival time difference, degree of enhancement, distribution uniformity of contrast medium, presence of non-enhancing region, and arterial filling mode in the lesion. A p < 0.05 was considered statistically significant. Results: Presence of air bronchial sign (odds ratio [OR] = 6.18, p = 0.025), acute angle between the lesion border and thoracic wall (OR = 7.124, p = 0.033), and homogeneous enhancement (OR = 35.26, p = 0.01) showed predictive value for the diagnosis of FOP. According to the results of the logistic regression analysis, the area under the receiver operating curve of the above features combined was 0.960, and the sensitivity and specificity were 95.0% and 82.6%, respectively. Conclusions: US combined with CEUS has the potential to differentiate between FOP and PLM.

Breast tubular adenomas are rare benign breast tumors and detailed descriptions of their sonographic appearance are necessary for differential diagnosis from fibroadenomas or breast cancers. This study investigated twenty-one histology-proved tubular adenomas in 17 patients and also included 48 fibroadenomas in 35 patients as a control group. There was no significant difference between the two groups with clinical presentation, which was age, tumor location, tumor number (p>0.05). Statistic analysis showed three significant factors in the differential diagnosis of tubular adenomas and fibroadenomas, including macro-lobulation (p=0.01), "tiny branch like" patterns (p=0.001) and vascularity (p=0.02). Other ultrasonographic features such as echogenicity, border, uniformity of echotexture, posterior acoustic enhancement, lateral wall shadowing were of no clinical significance (p>0.05). Calcifications were seen in three tubular adenomas which were different from those of carcinomas. Although tubular adenomas have some typical characteristics on sonography, surgery and core needle biopsy are still needed for complex cases to exclude progress to malignancy.

Background and Objectives: In this study, we assessed the utility of ultrasonography in monitoring the chemotherapeutic effects on primary thyroid lymphoma (PTL). Materials and Methods: This retrospective analysis included 17 patients with PTL who received chemotherapy from 2012 to 2022. The sonographic features were examined pre- and post-treatment using ultrasound (US) to monitor the treatment response at the first to second, third to fourth, and end cycles of chemotherapy and follow-up, and progression-free survival (PFS) and overall survival (OS) were analyzed. Results: The sonographic findings for all the patients indicated diffuse or nodular infiltration with markedly hypoechoic masses, and “stripe-shaped” high echoes and posterior acoustic enhancement were observed. Following one to two cycles of chemotherapy, a US examination revealed varying tumor reduction degrees and diminished blood flow signals. After three to four cycles of chemotherapy, the US demonstrated an evaluation efficacy comparable to that of PET-CT in cases in which the lesion had entirely disappeared postchemotherapy; however, its ability to differentiate between treatment response and residual lesions was less effective compared to that of PET-CT. After the end cycle of chemotherapy, the lesion sizes had significantly decreased compared to those at the baseline (p < 0.05). Postchemotherapy, Adler’s blood flow grades decreased significantly, with 80% graded as 0–1. Among the 10 patients with cervical lymph node enlargement, 70% showed reduced lesion sizes and blood flow signals. The cumulative 5-year PFS and OS rates were both 80% for the diffuse type and 82.5% and 78.8% for the nodular type, respectively (p > 0.05). Conclusions: US can be utilized to monitor the therapeutic response following chemotherapy for PTL, especially for early assessment and repeated dynamic monitoring, and can serve as a complementary follow-up method to PET-CT.

Background and Objectives: To evaluate the value of the postvascular phase of contrast-enhanced ultrasound (CEUS) in differentiating between benign and metastatic lymph nodes (LNs) in patients with breast cancer (BC). Materials and Methods: This study retrospectively analyzed 96 suspicious LNs in the lymphatic drainage area of the breast from 90 patients with BC. All LNs were assessed by conventional ultrasound (US) and CEUS following intravenous Sonazoid injection. All LNs underwent puncture biopsy, and pathological results were obtained. The correlations between US and CEUS indicators of LNs and LN metastasis (LNM) were analyzed. Results: Of the 96 LNs, 66 were metastatic. Overall, 80.00% (24/30) of the benign LNs exhibited relative hyper-enhancement in the postvascular phase, whereas 96.97% (64/66) of the metastatic LNs exhibited relative hypo-enhancement (p < 0.001). This CEUS finding was highly predictive of metastasis, with a sensitivity of 96.97%, specificity of 80.00%, positive predictive value of 91.43%, negative predictive value of 92.31%, and accuracy of 91.67%. The mean postvascular phase intensity (MPI) was significantly lower for malignant (median MPI, 12 dB) than for benign (median MPI, 75 dB) LNs. The postvascular phase was more sensitive, specific, and accurate than conventional US or the vascular phase of CEUS for the diagnosis of LNM, with an area under the curve of 0.95 (95% confidence interval: 0.89–0.99). Conclusions: Qualitative and quantitative indicators of the postvascular phase of CEUS provide a reliable diagnostic approach to differentiate benign and metastatic LNs in patients with BC.