Using time-series chest radiographs and laboratory data by machine learning for identifying pulmonary infection and colonization of Acinetobacter baumannii.

BACKGROUND: Accurately distinguishing between pulmonary infection and colonization in patients with Acinetobacter baumannii is of utmost importance to optimize treatment and prevent antibiotic abuse or inadequate therapy. An efficient automated sorting tool could prompt individualized interventions and enhance overall patient outcomes. This study aims to develop a robust machine learning classification model using a combination of time-series chest radiographs and laboratory data to accurately classify pulmonary status caused by Acinetobacter baumannii. METHODS: We proposed nested logistic regression models based on different time-series data to automatically classify the pulmonary status of patients with Acinetobacter baumannii. Advanced features were extracted from the time-series data of hospitalized patients, encompassing dynamic pneumonia indicators observed on chest radiographs and laboratory indicator values recorded at three specific time points. RESULTS: Data of 152 patients with Acinetobacter baumannii cultured from sputum or alveolar lavage fluid were retrospectively analyzed. Our model with multiple time-series data demonstrated a higher performance of AUC (0.850, with a 95% confidence interval of [0.638-0.873]), an accuracy of 0.761, a sensitivity of 0.833. The model, which only incorporated a single time point feature, achieved an AUC of 0.741. The influential model variables included difference in the chest radiograph pneumonia score. CONCLUSION: Dynamic assessment of time-series chest radiographs and laboratory data using machine learning allowed for accurate classification of colonization and infection with Acinetobacter baumannii. This demonstrates the potential to help clinicians provide individualized treatment through early detection.

Imaging diagnosis and differential diagnosis of extraskeletal osteosarcoma.

OBJECTIVE: The aim of this study was to investigate the clinical, imaging and pathological features of extraskeletal osteosarcoma (EOS) and to improve the understanding of this disease and other similar lesions. METHODS: The data for 11 patients with pathologically confirmed extraosseous osteosarcoma, including tumour site and size and imaging and clinical manifestations, were analysed retrospectively. RESULTS: Six patients were male (60%), and 5 were female (40%); patient age ranged from 23 to 76 years (average age 47.1 years). Among the 11 patients, 7 had clear calcifications or ossification with different morphologies, and 2 patients showed a massive mature bone tumour. MRI showed a mixed-signal mass with slightly longer T1 and T2 signals in the tumour parenchyma. Enhanced CT and MRI scans showed enhancement in the parenchyma. Ten patients had different degrees of necrosis and cystic degeneration in the mass, 2 of whom were complicated with haemorrhage, and MRI showed "fluid‒fluid level" signs. Of the 11 patients, five patients survived after surgery, and no obvious recurrence or metastasis was found on imaging examination. One patient died of lung metastasis after surgery, and 2 patients with open biopsy died of disease progression. One patient died of respiratory failure 2 months after operation. 2 patients had positive surgical margins, and 1 had lung metastasis 6 months after operation and died 19 months after operation. Another patient had recurrence 2 months after surgery. CONCLUSION: The diagnosis of EOS requires a combination of clinical, imaging and histological examinations. Cystic degeneration and necrosis; mineralization is common, especially thick and lumpy mineralization. Extended resection is still the first choice for localized lesions. For patients with positive surgical margins or metastases, adjuvant chemoradiotherapy is needed.

A deep learning-machine learning fusion approach for the classification of benign, malignant, and intermediate bone tumors.

OBJECTIVES: To build and validate deep learning and machine learning fusion models to classify benign, malignant, and intermediate bone tumors based on patient clinical characteristics and conventional radiographs of the lesion. METHODS: In this retrospective study, data were collected with pathologically confirmed diagnoses of bone tumors between 2012 and 2019. Deep learning and machine learning fusion models were built to classify tumors as benign, malignant, or intermediate using conventional radiographs of the lesion and potentially relevant clinical data. Five radiologists compared diagnostic performance with and without the model. Diagnostic performance was evaluated using the area under the curve (AUC). RESULTS: A total of 643 patients' (median age, 21 years; interquartile range, 12-38 years; 244 women) 982 radiographs were included. In the test set, the binary category classification task, the radiological model of classification for benign/not benign, malignant/nonmalignant, and intermediate/not intermediate had AUCs of 0.846, 0.827, and 0.820, respectively; the fusion models had an AUC of 0.898, 0.894, and 0.865, respectively. In the three-category classification task, the radiological model achieved a macro average AUC of 0.813, and the fusion model had a macro average AUC of 0.872. In the observation test, the mean macro average AUC of all radiologists was 0.819. With the three-category classification fusion model support, the macro AUC improved by 0.026. CONCLUSION: We built, validated, and tested deep learning and machine learning models that classified bone tumors at a level comparable with that of senior radiologists. Model assistance may somewhat help radiologists' differential diagnoses of bone tumors. KEY POINTS: • The deep learning model can be used to classify benign, malignant, and intermediate bone tumors. • The machine learning model fusing information from radiographs and clinical characteristics can improve the classification capacity for bone tumors. • The diagnostic performance of the fusion model is comparable with that of senior radiologists and is potentially useful as a complement to radiologists in a bone tumor differential diagnosis.

Clinical and imaging features preoperative evaluation of histological grade and microvascular infiltration of hepatocellular carcinoma.

BACKGROUND: To predict the histological grade and microvascular invasion (MVI) in patients with HCC. METHODS: A retrospective analysis was conducted on 175 patients who underwent MRI enhancement scanning (from September 2016.9 to October 2020). They were divided into MVI positive, MVI negative, Grade-high and Grade-low groups. RESULTS: The AFP of 175 HCC patients distributed in MVI positive and negative groups, Grade-low and Grade-high groups were statistically significant (P = 0.002 and 0.03, respectively). Multiple HCC lesions were more common in MVI positive and Grade-high groups. Correspondingly, more single lesions were found in MVI negative and Grade-low groups (P = 0.005 and 0.019, respectively). Capsule on MRI was more common in MVI negative and Grade-high groups, and the difference was statistically significant (P = 0.02 and 0.011, respectively). There were statistical differences in the distribution of three MRI signs: artistic rim enhancement, artistic peripheral enhancement, and tumor margin between MVI positive and MVI negative groups (P = 0.001, < 0.001, and < 0.001, respectively). Tumor hypointensity on HBP was significantly different between MVI positive and negative groups (P < 0.001). CONCLUSIONS: Our research shows that preoperative enhanced imaging can be used to predict MVI and tumor differentiation grade of HCC. The prognosis of MVI-negative group was better than that of MVI-positive group.

CT, MRI, and 18F-FDG PET/CT imaging features of seven cases of adult pancreatoblastoma.

OBJECTIVE: This study mainly analysed the imaging data for seven cases of adult pancreatoblastoma (PB) and summarized additional imaging features of this disease based on a literature review, aiming to improve the understanding and diagnosis rate of this disease. MATERIALS AND METHODS: The imaging data for seven adult patients pathologically diagnosed with adult PB were retrospectively analysed. Among the seven patients, six underwent computed tomography (CT) scans, two patients underwent abdominal magnetic resonance imaging (MRI), and five patients underwent 18F-FDG PET/CT. RESULTS: The tumours were located in the head of the pancreas in three cases, in the tail of the pancreas in two cases, and in the gastric antrum and neck of the pancreas in one case. Six tumours showed blurred edges, and an incomplete envelope was observed in only two cases when enhanced, which showed extruded growth and cyst-solid masses; one tumour was a solid mass with ossification. Showing mild or significant enhancement in the arterial phase (AP) for six cases. In the MRI sequence, isointensity was found on suppressed T1-weighted imaging, and hyperintensity was noted on suppressed T2-weighted imaging in two cases, with significant enhancement. Pancreatic duct dilatation was found in four cases. Tumour 18F-FDG PET/CT imaging exhibited high uptake in five cases. CONCLUSION: Adult PB involves a single tumour and commonly manifests as cystic-solid masses with blurred edges. Capsules are rare, ossification is an important feature, tumours can also present in ectopic pancreatic tissues, with mild or strengthening in the AP, and 18F-FDG uptake is high. These features are relatively specific characteristics in adult PB.