Background: This study aimed to evaluate the applicability of deep learning technology to thyroid ultrasound images for classification of thyroid nodules. Methods: This retrospective analysis included ultrasound images of patients with thyroid nodules investigated by fine-needle aspiration at the thyroid clinic of a single center from April 2010 to September 2012. Thyroid nodules with cytopathologic results of Bethesda category V (suspicious for malignancy) or VI (malignant) were defined as thyroid cancer. Multiple deep learning algorithms based on convolutional neural networks (CNNs) —ResNet, DenseNet, and EfficientNet—were utilized, and Siamese neural networks facilitated multi-view analysis of paired transverse and longitudinal ultrasound images. Results: Among 1,048 analyzed thyroid nodules from 943 patients, 306 (29%) were identified as thyroid cancer. In a subgroup analysis of transverse and longitudinal images, longitudinal images showed superior prediction ability. Multi-view modeling, based on paired transverse and longitudinal images, significantly improved the model performance; with an accuracy of 0.82 (95% confidence intervals [CI], 0.80 to 0.86) with ResNet50, 0.83 (95% CI, 0.83 to 0.88) with DenseNet201, and 0.81 (95% CI, 0.79 to 0.84) with EfficientNetv2_ s. Training with high-resolution images obtained using the latest equipment tended to improve model performance in association with increased sensitivity. Conclusion CNN algorithms applied to ultrasound images demonstrated substantial accuracy in thyroid nodule classification, indicating their potential as valuable tools for diagnosing thyroid cancer. However, in real-world clinical settings, it is important to aware that model performance may vary depending on the quality of images acquired by different physicians and imaging devices.

Purpose: Pulmonary tumor thrombotic microangiopathy (PTTM) is a fatal complication of gastric cancer (GC). This study aimed to evaluate the clinical characteristics, outcomes, and immunohistochemical profiles of patients with GC-induced PTTM. Materials and Methods: From 2011 to 2023, 8 patients were clinically diagnosed with PTTM associated with GC antemortem. Clinical features and outcomes were reviewed, and immunohistochemical staining for c-erbB-2, MutL protein homolog 1, and programmed cell death ligand-1 was performed. Results: The median patient age was 56 years (range, 34–66 years). In all the patients, the tumors exhibited either ulceroinfiltrative or diffusely infiltrative gross morphology.The median tumor size was 5.8 cm (range, 2.0 cm–15.0 cm). Poorly differentiated adenocarcinoma was the most common histological type (6/8, 75%), followed by signet ring cell carcinoma (1/8, 12.5%) and moderately differentiated adenocarcinoma (1/8, 12.5%).Chest computed tomography revealed ground-glass opacities (7/8, 87.5%) or tree-in-bud signs (2/8, 25.0%) without definite evidence of pulmonary thromboembolism. Disseminated intravascular coagulation was present in 62.5% (5/8) of the patients diagnosed with PTTM.C-erbB-2 was positive in one patient (1/8, 12.5%). One patient who received palliative chemotherapy after developing PTTM survived for 35 days, whereas the other 7 patients who did not receive chemotherapy after developing PTTM survived for 7 days or less after PTTM diagnosis. Conclusions Most patients with GC-induced PTTM had an undifferentiated-type histology, infiltrative morphology, and extremely poor survival. Palliative chemotherapy may benefit patients with GC-induced PTTM; however, further studies are needed to explore the potential of targeted therapy in these patients.

Background: This study aimed to evaluate the applicability of deep learning technology to thyroid ultrasound images for classification of thyroid nodules. Methods: This retrospective analysis included ultrasound images of patients with thyroid nodules investigated by fine-needle aspiration at the thyroid clinic of a single center from April 2010 to September 2012. Thyroid nodules with cytopathologic results of Bethesda category V (suspicious for malignancy) or VI (malignant) were defined as thyroid cancer. Multiple deep learning algorithms based on convolutional neural networks (CNNs) —ResNet, DenseNet, and EfficientNet—were utilized, and Siamese neural networks facilitated multi-view analysis of paired transverse and longitudinal ultrasound images. Results: Among 1,048 analyzed thyroid nodules from 943 patients, 306 (29%) were identified as thyroid cancer. In a subgroup analysis of transverse and longitudinal images, longitudinal images showed superior prediction ability. Multi-view modeling, based on paired transverse and longitudinal images, significantly improved the model performance; with an accuracy of 0.82 (95% confidence intervals [CI], 0.80 to 0.86) with ResNet50, 0.83 (95% CI, 0.83 to 0.88) with DenseNet201, and 0.81 (95% CI, 0.79 to 0.84) with EfficientNetv2_ s. Training with high-resolution images obtained using the latest equipment tended to improve model performance in association with increased sensitivity. Conclusion CNN algorithms applied to ultrasound images demonstrated substantial accuracy in thyroid nodule classification, indicating their potential as valuable tools for diagnosing thyroid cancer. However, in real-world clinical settings, it is important to aware that model performance may vary depending on the quality of images acquired by different physicians and imaging devices.

Background: This study aimed to evaluate the applicability of deep learning technology to thyroid ultrasound images for classification of thyroid nodules. Methods: This retrospective analysis included ultrasound images of patients with thyroid nodules investigated by fine-needle aspiration at the thyroid clinic of a single center from April 2010 to September 2012. Thyroid nodules with cytopathologic results of Bethesda category V (suspicious for malignancy) or VI (malignant) were defined as thyroid cancer. Multiple deep learning algorithms based on convolutional neural networks (CNNs) —ResNet, DenseNet, and EfficientNet—were utilized, and Siamese neural networks facilitated multi-view analysis of paired transverse and longitudinal ultrasound images. Results: Among 1,048 analyzed thyroid nodules from 943 patients, 306 (29%) were identified as thyroid cancer. In a subgroup analysis of transverse and longitudinal images, longitudinal images showed superior prediction ability. Multi-view modeling, based on paired transverse and longitudinal images, significantly improved the model performance; with an accuracy of 0.82 (95% confidence intervals [CI], 0.80 to 0.86) with ResNet50, 0.83 (95% CI, 0.83 to 0.88) with DenseNet201, and 0.81 (95% CI, 0.79 to 0.84) with EfficientNetv2_ s. Training with high-resolution images obtained using the latest equipment tended to improve model performance in association with increased sensitivity. Conclusion CNN algorithms applied to ultrasound images demonstrated substantial accuracy in thyroid nodule classification, indicating their potential as valuable tools for diagnosing thyroid cancer. However, in real-world clinical settings, it is important to aware that model performance may vary depending on the quality of images acquired by different physicians and imaging devices.

Purpose: Pulmonary tumor thrombotic microangiopathy (PTTM) is a fatal complication of gastric cancer (GC). This study aimed to evaluate the clinical characteristics, outcomes, and immunohistochemical profiles of patients with GC-induced PTTM. Materials and Methods: From 2011 to 2023, 8 patients were clinically diagnosed with PTTM associated with GC antemortem. Clinical features and outcomes were reviewed, and immunohistochemical staining for c-erbB-2, MutL protein homolog 1, and programmed cell death ligand-1 was performed. Results: The median patient age was 56 years (range, 34–66 years). In all the patients, the tumors exhibited either ulceroinfiltrative or diffusely infiltrative gross morphology.The median tumor size was 5.8 cm (range, 2.0 cm–15.0 cm). Poorly differentiated adenocarcinoma was the most common histological type (6/8, 75%), followed by signet ring cell carcinoma (1/8, 12.5%) and moderately differentiated adenocarcinoma (1/8, 12.5%).Chest computed tomography revealed ground-glass opacities (7/8, 87.5%) or tree-in-bud signs (2/8, 25.0%) without definite evidence of pulmonary thromboembolism. Disseminated intravascular coagulation was present in 62.5% (5/8) of the patients diagnosed with PTTM.C-erbB-2 was positive in one patient (1/8, 12.5%). One patient who received palliative chemotherapy after developing PTTM survived for 35 days, whereas the other 7 patients who did not receive chemotherapy after developing PTTM survived for 7 days or less after PTTM diagnosis. Conclusions Most patients with GC-induced PTTM had an undifferentiated-type histology, infiltrative morphology, and extremely poor survival. Palliative chemotherapy may benefit patients with GC-induced PTTM; however, further studies are needed to explore the potential of targeted therapy in these patients.

Background: This study aimed to evaluate the applicability of deep learning technology to thyroid ultrasound images for classification of thyroid nodules. Methods: This retrospective analysis included ultrasound images of patients with thyroid nodules investigated by fine-needle aspiration at the thyroid clinic of a single center from April 2010 to September 2012. Thyroid nodules with cytopathologic results of Bethesda category V (suspicious for malignancy) or VI (malignant) were defined as thyroid cancer. Multiple deep learning algorithms based on convolutional neural networks (CNNs) —ResNet, DenseNet, and EfficientNet—were utilized, and Siamese neural networks facilitated multi-view analysis of paired transverse and longitudinal ultrasound images. Results: Among 1,048 analyzed thyroid nodules from 943 patients, 306 (29%) were identified as thyroid cancer. In a subgroup analysis of transverse and longitudinal images, longitudinal images showed superior prediction ability. Multi-view modeling, based on paired transverse and longitudinal images, significantly improved the model performance; with an accuracy of 0.82 (95% confidence intervals [CI], 0.80 to 0.86) with ResNet50, 0.83 (95% CI, 0.83 to 0.88) with DenseNet201, and 0.81 (95% CI, 0.79 to 0.84) with EfficientNetv2_ s. Training with high-resolution images obtained using the latest equipment tended to improve model performance in association with increased sensitivity. Conclusion CNN algorithms applied to ultrasound images demonstrated substantial accuracy in thyroid nodule classification, indicating their potential as valuable tools for diagnosing thyroid cancer. However, in real-world clinical settings, it is important to aware that model performance may vary depending on the quality of images acquired by different physicians and imaging devices.

Purpose: Pulmonary tumor thrombotic microangiopathy (PTTM) is a fatal complication of gastric cancer (GC). This study aimed to evaluate the clinical characteristics, outcomes, and immunohistochemical profiles of patients with GC-induced PTTM. Materials and Methods: From 2011 to 2023, 8 patients were clinically diagnosed with PTTM associated with GC antemortem. Clinical features and outcomes were reviewed, and immunohistochemical staining for c-erbB-2, MutL protein homolog 1, and programmed cell death ligand-1 was performed. Results: The median patient age was 56 years (range, 34–66 years). In all the patients, the tumors exhibited either ulceroinfiltrative or diffusely infiltrative gross morphology.The median tumor size was 5.8 cm (range, 2.0 cm–15.0 cm). Poorly differentiated adenocarcinoma was the most common histological type (6/8, 75%), followed by signet ring cell carcinoma (1/8, 12.5%) and moderately differentiated adenocarcinoma (1/8, 12.5%).Chest computed tomography revealed ground-glass opacities (7/8, 87.5%) or tree-in-bud signs (2/8, 25.0%) without definite evidence of pulmonary thromboembolism. Disseminated intravascular coagulation was present in 62.5% (5/8) of the patients diagnosed with PTTM.C-erbB-2 was positive in one patient (1/8, 12.5%). One patient who received palliative chemotherapy after developing PTTM survived for 35 days, whereas the other 7 patients who did not receive chemotherapy after developing PTTM survived for 7 days or less after PTTM diagnosis. Conclusions Most patients with GC-induced PTTM had an undifferentiated-type histology, infiltrative morphology, and extremely poor survival. Palliative chemotherapy may benefit patients with GC-induced PTTM; however, further studies are needed to explore the potential of targeted therapy in these patients.

Objective: To investigate the incidence and timing of late growth of subsolid nodules (SSNs) ≥6 mm after initial 5-year stability, its clinical implications, and the appropriate follow-up strategy. Materials and Methods: This retrospective study included SSNs ≥6 mm that remained stable for the initial five years after detection. The incidence and timing of subsequent growth after five years of stability were analyzed using the Kaplan–Meier method. Descriptive analyses were conducted to evaluate the clinical stage shift in the SSNs, showing growth and the presence of metastasis during the follow-up period. Finally, an effective follow-up CT scan strategy for managing SSNs after a 5-year period of stability was investigated. Results: Two hundred thirty-five eligible SSNs (211 pure ground-glass and 24 part-solid nodules) in 235 patients (median age, 63 years; 132 female) were followed for additional <1 to 181 months (median, 87.0 months; interquartile range [IQR], 47.0– 119.0 months) after 5-year stability. Fourteen SSNs (6.0%) showed growth at two to 145 months (median, 96 months; IQR:43.0–122.25 months) from the CT scan confirming 5-year stability, with the estimated cumulative incidence of growth of 0.4%, 2.1%, and 6.5% at 1, 5, and 10 years, respectively. Nine SSNs (3.8%) exhibited clinical stage shifts. No lung cancer metastases were observed. Hypothetical follow-up CT scans performed at 5, 10, and 15 years after 5-years of stability, would have detected 5 (36%), 11 (79%), and 14 (100%) of the 14 growing SSNs, along with 4 (44%), 8 (89%), and 9 (100%) of the nine stage shifts, respectively. Conclusion During a long-term follow-up of pulmonary SSNs ≥6 mm after 5-years of stability, a low incidence of growth without occurrence of metastasis was noted. CT scans every five years after the initial 5-year stability period may be reasonable.

Objective: To investigate the incidence and timing of late growth of subsolid nodules (SSNs) ≥6 mm after initial 5-year stability, its clinical implications, and the appropriate follow-up strategy. Materials and Methods: This retrospective study included SSNs ≥6 mm that remained stable for the initial five years after detection. The incidence and timing of subsequent growth after five years of stability were analyzed using the Kaplan–Meier method. Descriptive analyses were conducted to evaluate the clinical stage shift in the SSNs, showing growth and the presence of metastasis during the follow-up period. Finally, an effective follow-up CT scan strategy for managing SSNs after a 5-year period of stability was investigated. Results: Two hundred thirty-five eligible SSNs (211 pure ground-glass and 24 part-solid nodules) in 235 patients (median age, 63 years; 132 female) were followed for additional <1 to 181 months (median, 87.0 months; interquartile range [IQR], 47.0– 119.0 months) after 5-year stability. Fourteen SSNs (6.0%) showed growth at two to 145 months (median, 96 months; IQR:43.0–122.25 months) from the CT scan confirming 5-year stability, with the estimated cumulative incidence of growth of 0.4%, 2.1%, and 6.5% at 1, 5, and 10 years, respectively. Nine SSNs (3.8%) exhibited clinical stage shifts. No lung cancer metastases were observed. Hypothetical follow-up CT scans performed at 5, 10, and 15 years after 5-years of stability, would have detected 5 (36%), 11 (79%), and 14 (100%) of the 14 growing SSNs, along with 4 (44%), 8 (89%), and 9 (100%) of the nine stage shifts, respectively. Conclusion During a long-term follow-up of pulmonary SSNs ≥6 mm after 5-years of stability, a low incidence of growth without occurrence of metastasis was noted. CT scans every five years after the initial 5-year stability period may be reasonable.

Objective: To investigate the incidence and timing of late growth of subsolid nodules (SSNs) ≥6 mm after initial 5-year stability, its clinical implications, and the appropriate follow-up strategy. Materials and Methods: This retrospective study included SSNs ≥6 mm that remained stable for the initial five years after detection. The incidence and timing of subsequent growth after five years of stability were analyzed using the Kaplan–Meier method. Descriptive analyses were conducted to evaluate the clinical stage shift in the SSNs, showing growth and the presence of metastasis during the follow-up period. Finally, an effective follow-up CT scan strategy for managing SSNs after a 5-year period of stability was investigated. Results: Two hundred thirty-five eligible SSNs (211 pure ground-glass and 24 part-solid nodules) in 235 patients (median age, 63 years; 132 female) were followed for additional <1 to 181 months (median, 87.0 months; interquartile range [IQR], 47.0– 119.0 months) after 5-year stability. Fourteen SSNs (6.0%) showed growth at two to 145 months (median, 96 months; IQR:43.0–122.25 months) from the CT scan confirming 5-year stability, with the estimated cumulative incidence of growth of 0.4%, 2.1%, and 6.5% at 1, 5, and 10 years, respectively. Nine SSNs (3.8%) exhibited clinical stage shifts. No lung cancer metastases were observed. Hypothetical follow-up CT scans performed at 5, 10, and 15 years after 5-years of stability, would have detected 5 (36%), 11 (79%), and 14 (100%) of the 14 growing SSNs, along with 4 (44%), 8 (89%), and 9 (100%) of the nine stage shifts, respectively. Conclusion During a long-term follow-up of pulmonary SSNs ≥6 mm after 5-years of stability, a low incidence of growth without occurrence of metastasis was noted. CT scans every five years after the initial 5-year stability period may be reasonable.