Sriwijaya Journal of Radiology and Imaging Research

Loculated Right-Sided Hydropneumothorax Mimicking Giant Pulmonary Bullae in a Post-Tuberculosis Patient: A Multimodality Imaging Diagnostic Challenge

2026-07-16T03:46:43+00:00

Introduction: Post-tuberculosis lung disease remains a significant public health challenge affecting millions of individuals globally, representing a substantial health burden in tuberculosis-endemic regions and in developed countries with immigration from endemic areas. Loculated hydropneumothorax as a late complication of successfully treated pulmonary tuberculosis is a rare but diagnostically challenging entity, particularly when imaging findings suggest alternative pathology such as giant pulmonary bullae. This case illustrates the complexity of post-tuberculosis complications and the essential role of multimodality imaging.

Case Presentation: A 63-year-old retired woman presented to the emergency department with three days of progressive dyspnea accompanied by productive cough with yellowish-white sputum. Physical examination revealed severe tachypnea (41 breaths per minute), clinically significant hypoxemia (SpO₂ 88 percent on room air), and diminished breath sounds over the right hemithorax with crackles in the right upper lobe. Chest radiography demonstrated a large thin-walled cavity (18 by 9.5 by 14 centimeters) with a horizontal air-fluid level in the right hemithorax, mediastinal leftward shift, and right costophrenic sinus obliteration. Thoracic point-of-care ultrasound revealed predominant gas throughout the right hemithorax with minimal pleural fluid in dependent zones and absence of identifiable lung tissue above the hemidiaphragm. Contrast-enhanced computed tomography definitively identified loculated right-sided hydropneumothorax with thin-walled pleural compartment, air-fluid level, compressive atelectasis of right lower and middle lobes, and post-tuberculosis fibrotic sequelae. Individual imaging modalities — radiography, ultrasound, and computed tomography — each contributed essential diagnostic information, demonstrating that none is sufficient in isolation.

Conclusion: Loculated hydropneumothorax must be considered in the differential diagnosis of large cavitary lesions in post-tuberculosis patients. A multimodality imaging approach is essential for achieving diagnostic certainty and preventing unnecessary surgical intervention.

Diagnostic Accuracy of Multiparametric MRI-Based Machine-Learning Radiomics for Differentiating Malignant from Benign Soft-Tissue Tumours: A Multi-Institutional Study

2026-07-20T04:45:29+00:00

Introduction: Reliable preoperative discrimination of malignant from benign soft-tissue tumours (STTs) governs biopsy, surgical-margin and neoadjuvant decisions, yet conventional MRI interpretation is experience-dependent and biopsy is invasive and prone to sampling error. We aimed to develop and internally validate a machine-learning radiomics model from multiparametric MRI (mpMRI) for this task across multiple institutions.

Methods: In this STARD 2015-compliant retrospective multi-institutional diagnostic-accuracy study, 215 patients (132 benign, 83 malignant) with histopathologically confirmed STTs imaged at three South Sumatran centres (2019–2023) were split 70:30 into training (n=150) and validation (n=65) cohorts. Radiomic features from T1W, T2W fat-suppressed and ADC maps underwent ICC-stability filtering and LASSO selection; SVM, Random Forest and XGBoost classifiers were compared against histopathology (reference standard) and blinded radiologist visual reads. Sensitivity, specificity, predictive values, likelihood ratios (95% CIs), ROC (DeLong), Cohen’s κ and multivariable logistic regression were computed.

Results: A 14-feature signature was selected from 945 ICC-stable features. In internal validation, XGBoost achieved AUC 0.92 (95% CI 0.88–0.95), sensitivity 86.7% (70.3–94.7), specificity 91.4% (77.6–97.0), PPV 89.7%, NPV 88.9%, accuracy 89.2%, LR+ 10.1 and LR− 0.15. XGBoost exceeded SVM (AUC 0.84; DeLong p=0.012) and radiologist visual read (AUC 0.78; p<0.001; McNemar p=0.027). Inter-reader κ was 0.78 (0.63–0.94). The radiomics signature (adjusted OR 3.32, p<0.001) and lower ADC (OR 0.21, p<0.001) were independent malignancy predictors.

Conclusion: An mpMRI XGBoost radiomics model provides accurate, non-invasive discrimination of malignant from benign STTs with high specificity and a clinically useful positive likelihood ratio, supporting its role as PACS-integrated decision support and as a triage tool in resource-variable settings.

Vision Transformer Reconstruction for Super-Resolution and Artifact Reduction in 1.5-Tesla Fast-Spin-Echo Pelvic MRI: A Diagnostic-Accuracy Study

2026-08-07T02:36:57+00:00

Introduction: Fast-spin-echo (FSE) MRI is the reference modality for pelvic evaluation, but high-resolution acquisition is slow and motion-prone. Deep-learning reconstruction may recover diagnostic quality from short, motion-tolerant acquisitions, yet most evidence relies on image-similarity indices rather than radiologist performance. We validated a Vision-Transformer (ViT) reconstruction for simultaneous super-resolution and motion-artifact reduction in 1.5-Tesla T2-weighted FSE pelvic MRI.

Methods: In this retrospective diagnostic-accuracy study (STARD 2015) at a tertiary hospital in Palembang, Indonesia, 450 examinations (development n=360; test n=90) were analysed. A U-shaped shifted-window ViT reconstructed high-resolution images from retrospectively degraded low-resolution/motion-corrupted inputs. Two blinded radiologists scored each test case under native low-resolution, UNet- and ViT-reconstructed conditions against a histopathology/expert-consensus reference standard. Sensitivity, specificity, predictive values, AUC, likelihood ratios (95% CI), inter-reader kappa, DeLong and McNemar tests, and multivariable logistic regression were computed.

Results: Target prevalence was 53.3%. ViT reconstruction achieved sensitivity 93.8% (95% CI 83.2–97.9), specificity 88.1% (75.0–94.8), AUC 0.943 (0.894–0.992), LR+ 7.87 and LR− 0.071, versus AUC 0.881 (UNet) and 0.751 (low-resolution); ViT vs low-resolution DeLong p<0.001, McNemar p<0.001. Inter-reader agreement rose from kappa 0.49 to 0.87. ViT gave the best fidelity (PSNR 34.82 dB; SSIM 0.941; p<0.001 vs UNet) at 0.15 s/slice. Sub-centimetre lesions (OR 3.84, p=0.006) and severe motion (OR 2.97, p=0.029) independently predicted error.

Conclusion: A shifted-window Vision Transformer recovered diagnostic-quality pelvic FSE MRI from short, motion-tolerant acquisitions, significantly improving radiologist lesion detection and inter-reader agreement over convolutional reconstruction. The real-time, PACS-compatible pipeline is promising for high-throughput pelvic MRI and warrants prospective validation.

Intra-individual Comparison of High-Relaxivity versus Standard Macrocyclic Gadolinium Agents for Detecting Hepatic Micrometastases at 3.0 T

2026-08-10T03:18:02+00:00

Introduction. Hepatic micrometastases smaller than 10 mm critically influence oncologic staging and eligibility for curative metastasectomy, yet the sub-centimetre sensitivity of standard gadolinium-enhanced MRI is limited. High-relaxivity macrocyclic gadolinium-based contrast agents (GBCAs) generate higher lesion-to-liver contrast-to-noise ratio (CNR) and may improve detection. We compared a high-relaxivity versus a standard macrocyclic GBCA at equimolar dose within the same patients.

Methods. In this prospective intra-individual diagnostic-accuracy study reported per STARD 2015 at a tertiary hospital in Palembang, Indonesia, 48 adult oncology patients underwent two 3.0-T liver MRI examinations 2 to 7 days apart in randomised contrast order (standard gadoterate meglumine versus a high-relaxivity macrocyclic agent, 0.1 mmol/kg). Two gastrointestinal radiologists blinded to agent and clinical and reference data read images independently. A composite reference standard (histopathology and ≥6-month multiphasic imaging follow-up) defined 134 metastatic lesions. Sensitivity, specificity, AUC (DeLong), likelihood ratios, Cohen kappa and McNemar tests were computed.

Results. The high-relaxivity agent increased CNR by 12.2 units (48.5 versus 36.2; p<0.001). Per-lesion sensitivity rose to 94.8% (95% CI 89.6–97.4) from 82.8% (75.6–88.3), and micrometastasis sensitivity to 88.7% from 66.1% (McNemar p<0.001). The area under the ROC curve was 0.929 (0.893–0.966) versus 0.821 (0.757–0.886; DeLong p=0.003), and the negative likelihood ratio improved to 0.06. Inter-reader kappa was 0.89 versus 0.83. Specificity was comparable (84.5% versus 89.7%; p=0.41).

Conclusion. At equimolar dose, the high-relaxivity macrocyclic GBCA significantly improved CNR and sub-centimetre hepatic metastasis detection without meaningful loss of specificity, a gain that may alter oncologic management. High-relaxivity agents are recommended for high-risk hepatic staging.

Resting-State Functional MRI Connectivity Disruption Predicts Post-Stroke Epileptogenesis: A Prospective Longitudinal Cohort Study

2026-08-11T03:47:55+00:00

Introduction: Post-stroke epilepsy (PSE) complicates roughly 5–10% of ischaemic strokes, yet clinical and electroencephalographic markers predict unprovoked late seizures only modestly. Resting-state functional MRI (rs-fMRI) with graph theory can non-invasively quantify brain-network architecture. We tested whether subacute functional-connectivity disruption predicts PSE.

Methods: In a prospective longitudinal cohort at a tertiary hospital in Palembang, Indonesia, 150 adults with first-ever supratentorial ischaemic stroke underwent 3.0-T rs-fMRI on day 7–14 and were followed for 24 months (reported per STARD 2015 and TRIPOD). Automated Anatomical Labelling 90-region graph metrics were derived (CONN/SPM12). The reference standard was an International League Against Epilepsy-defined unprovoked late seizure, adjudicated blind to imaging. A penalised support-vector-machine model was internally validated (nested cross-validation, optimism correction, calibration) and compared with a clinical model using DeLong, decision-curve and competing-risks analyses.

Results: PSE occurred in 30 of 150 patients (cumulative incidence 19.2%). PSE patients showed thalamic degree-centrality overload (62.4±8.1 vs 45.2±6.8; p<0.001) and small-world collapse (σ 1.08±0.12 vs 1.25±0.11; p=0.008). The rs-fMRI model achieved sensitivity 86.7% (95% CI 70.3–94.7), specificity 88.3% (81.4–92.9), AUC 0.92 (0.85–0.99), LR+ 7.43 and LR− 0.15, versus clinical AUC 0.74 (DeLong p<0.001); inter-reader kappa was 0.84.

Conclusion: Subacute rs-fMRI connectomic disruption is a strong, independent, internally validated predictor of PSE that outperforms clinical variables. External multicentre validation is warranted before clinical adoption.