Lesion-conditioning of synthetic MRI-derived subtraction-MIPs of the breast using a latent diffusion model.

The purpose of this feasibility study is to investigate if latent diffusion models (LDMs) are capable to generate contrast enhanced (CE) MRI-derived subtraction maximum intensity projections (MIPs) of the breast, which are conditioned by lesions. We trained an LDM with n = 2832 CE-MIPs of breast MRI examinations of n = 1966 patients (median age: 50 years) acquired between the years 2015 and 2020. The LDM was subsequently conditioned with n = 756 segmented lesions from n = 407 examinations, indicating their location and BI-RADS scores. By applying the LDM, synthetic images were generated from the segmentations of an independent validation dataset. Lesions, anatomical correctness, and realistic impression of synthetic and real MIP images were further assessed in a multi-rater study with five independent raters, each evaluating n = 204 MIPs (50% real/50% synthetic images). The detection of synthetic MIPs by the raters was akin to random guessing with an AUC of 0.58. Interrater reliability of the lesion assessment was high both for real (Kendall's W = 0.77) and synthetic images (W = 0.85). A higher AUC was observed for the detection of suspicious lesions (BI-RADS ≥ 4) in synthetic MIPs (0.88 vs. 0.77; p = 0.051). Our results show that LDMs can generate lesion-conditioned MRI-derived CE subtraction MIPs of the breast, however, they also indicate that the LDM tended to generate rather typical or 'textbook representations' of lesions.

Free-Breathing Low-Field MRI of the Lungs Detects Functional Alterations Associated With Persistent Symptoms After COVID-19 Infection.

OBJECTIVES: With the COVID-19 pandemic, repetitive lung examinations have become necessary to follow-up symptoms and associated alterations. Low-field MRI, benefiting from reduced susceptibility effects, is a promising alternative for lung imaging to limit radiations absorbed by patients during CT examinations, which also have limited capability to assess functional alterations. The aim of this investigative study was to explore the functional abnormalities that free-breathing 0.55 T MRI in combination with the phase-resolved functional lung (PREFUL) analysis could identify in patients with persistent symptoms after COVID-19 infection. MATERIALS AND METHODS: Seventy-four COVID-19 patients and 8 healthy volunteers were prospectively scanned in free-breathing with a balanced steady-state free-precession sequence optimized at 0.55 T, 5 months postinfection on average. Normalized perfusion (Q), fractional ventilation (FV), and flow-volume loop correlation (FVLc) maps were extracted with the PREFUL technique. Q, FV, and FVLc defects as well as defect overlaps between these metrics were quantified. Morphological turbo-spin-echo images were also acquired, and the extent of abnormalities was scored by a board-certified radiologist. To investigate the functional correlates of persistent symptoms, a recursive feature elimination algorithm was applied to find the most informative variables to detect the presence of persistent symptoms with a logistic regression model and a cross-validation strategy. All MRI metrics, sex, age, body mass index, and the presence of preexisting lung conditions were included. RESULTS: The most informative variables to detect persistent symptoms were the percentage of concurrent Q and FVLc defects and of areas free of those defects. A detection accuracy of 71.4% was obtained with these 2 variables when fitting the model on the entire dataset. Although none of the single variables differed between patients with and without persistent symptoms ( P > 0.05), the combined score of these 2 variables did ( P < 0.02). This score also showed a consistent increase from healthy volunteers (7.7) to patients without persistent symptoms (8.2) and with persistent symptoms (8.6). The morphological abnormality score showed poor correlation with the functional parameters. CONCLUSIONS: Functional pulmonary examinations using free-breathing 0.55 T MRI with PREFUL analysis revealed potential quantitative markers of impaired lung function in patients with persistent symptoms after COVID-19 infection, potentially complementing morphologic imaging. Future work is needed to explore the translational relevance and clinical implication of these findings.