Lung nodule detection in chest X-rays using synthetic ground-truth data comparing CNN-based diagnosis to human performance.

We present a method to generate synthetic thorax radiographs with realistic nodules from CT scans, and a perfect ground truth knowledge. We evaluated the detection performance of nine radiologists and two convolutional neural networks in a reader study. Nodules were artificially inserted into the lung of a CT volume and synthetic radiographs were obtained by forward-projecting the volume. Hence, our framework allowed for a detailed evaluation of CAD systems' and radiologists' performance due to the availability of accurate ground-truth labels for nodules from synthetic data. Radiographs for network training (U-Net and RetinaNet) were generated from 855 CT scans of a public dataset. For the reader study, 201 radiographs were generated from 21 nodule-free CT scans with altering nodule positions, sizes and nodule counts of inserted nodules. Average true positive detections by nine radiologists were 248.8 nodules, 51.7 false positive predicted nodules and 121.2 false negative predicted nodules. The best performing CAD system achieved 268 true positives, 66 false positives and 102 false negatives. Corresponding weighted alternative free response operating characteristic figure-of-merits (wAFROC FOM) for the radiologists range from 0.54 to 0.87 compared to a value of 0.81 (CI 0.75-0.87) for the best performing CNN. The CNN did not perform significantly better against the combined average of the 9 readers (p = 0.49). Paramediastinal nodules accounted for most false positive and false negative detections by readers, which can be explained by the presence of more tissue in this area.

Detection of Bone Marrow Edema in Patients with Osteoid Osteoma Using Three-Material Decomposition with Dual-Layer Spectral CT.

The aim of this study is to assess whether perifocal bone marrow edema (BME) in patients with osteoid osteoma (OO) can be accurately detected on dual-layer spectral CT (DLCT) with three-material decomposition. To that end, 18 patients with OO (25.33 ± 12.44 years; 7 females) were pairwise-matched with 18 patients (26.72 ± 9.65 years; 9 females) admitted for suspected pathologies other than OO in the same anatomic location but negative imaging findings. All patients were examined with DLCT and MRI. DLCT data was decomposed into hydroxyapatite and water- and fat-equivalent volume fraction maps. Two radiologists assessed DLCT-based volume fraction maps for the presence of perifocal BME, using a Likert scale (1 = no edema; 2 = likely no edema; 3 = likely edema; 4 = edema). Accuracy, sensitivity, and specificity for the detection of BME on DLCT were analyzed using MR findings as standard of reference. For the detection of BME in patients with OO, DLCT showed a sensitivity of 0.92, a specificity of 0.94, and an accuracy of 0.92 for both radiologists. Interreader agreement for the assessment of BME with DLCT was substantial (weighted κ = 0.78; 95% CI, 0.59, 0.94). DLCT with material-specific volume fraction maps allowed accurate detection of BME in patients with OO. This may spare patients additional examinations and facilitate the diagnosis of OO.

Cartilage repair surgery prevents progression of knee degeneration.

PURPOSE: To investigate, whether cartilage repair surgery for focal osteochondral defects at the knee results in less degenerative changes over 6 years in a MR imaging follow-up than morphologically initially identical defects in non-operated control subjects from the osteoarthritis initiative (OAI). METHODS: A total of 32 individuals received baseline and follow-up MRI. In n = 16 patients with cartilage repair [osteochondral autograft transfer system (OATS), n = 12; spongiosa-augmented matrix-associated autologous chondrocyte implantation (MACI), n = 4] MRI was performed preoperatively and after 5.7 ± 2.3 year follow-up. Baseline MRIs of non-operated subjects from the OAI were screened for initially identical cartilage defects (n = 16). Morphological knee abnormalities were assessed using WORMS, AMADEUS and MOCART scores. A sagittal 2D MSME sequence was implemented for quantitative cartilage T2 relaxation time measurements in all (0, 2, 4, 6 and 8-years) follow-ups from the OAI and in the postoperative MRI protocol. RESULTS: For both groups, focal osteochondral defects were located at the femoral condyle in 8/16 cases (5 medial, 3 lateral) and at the patella in 8/16 cases. At baseline, the mean cartilage defect size ± SD was 1.4 ± 1.3 cm2 for the control group and 1.3 ± 1.2 cm2 for the cartilage repair group (n.s.). WORMS scores were not significantly different between the cartilage repair group and the control group at baseline [mean difference ± SEM (95%CI); 0.5 ± 2.5 (- 4.7, 5.7), n.s.]. During identical follow-up times, the progression of total WORMS scores [19.9 ± 2.3 (15.0, 24.9), P < 0.001] and of cartilage defects scores in the affected (P < 0.001) and in the opposing (P = 0.029) compartment was significantly more severe in non-operated individuals (P < 0.05). In non-operated subjects, T2 values increased continuously from baseline to the 8-year follow-up (P = 0.001). CONCLUSIONS: Patients with cartilage repair showed less progression of degenerative MRI changes at 6-year follow-up than a control cohort from the OAI with initially identical osteochondral defects. Patients with focal cartilage defects may profit from cartilage repair surgery since it may prevent progression of early osteoarthritis at the knee joint. LEVEL OF EVIDENCE: Prognostic study, Level II.

Association of diabetes mellitus and biochemical knee cartilage composition assessed by T2 relaxation time measurements: Data from the osteoarthritis initiative.

PURPOSE: To investigate the association of the presence and severity of diabetes mellitus (DM) with articular cartilage composition, using magnetic resonance imaging (MRI)-based T2 relaxation time measurements, and structural knee abnormalities. MATERIALS AND METHODS: In the Osteoarthritis Initiative 208, participants with DM (age 63.0 ± 8.9 years; 111 females) and risk factors for osteoarthritis (OA) or mild radiographic tibiofemoral OA (Kellgren-Lawrence [KL] grade ≤2) were identified and group-matched with 208 controls without DM (age 63.3 ± 9.1 years; 111 females). Subjects with diabetes-related renal or ophthalmological complications or insulin treatment at baseline (n = 50) were defined as severe DM. 3T MR images of the right knee were assessed for articular cartilage T2 , including texture and laminar analyses derived from the patella, medial, and lateral femur and tibia and for structural abnormalities using the modified whole-organ magnetic resonance imaging score (WORMS). Clustered linear regression analyses were used to assess associations of DM with MRI findings. RESULTS: DM subjects had significantly higher cartilage T2 in the patella (mean difference 0.92 msec [95% confidence interval (CI) 0.79, 1.06]; P = 0.001) and medial femur (mean difference 0.36 msec [95% CI 0.27, 0.81]; P = 0.006) compared to controls. Averaged over all compartments, DM subjects showed significantly higher texture parameters (variance, P = 0.001; contrast, P = 0.002; entropy, P < 0.001). Subjects with severe DM additionally showed higher T2 in the medial tibial deep and superficial layers (P = 0.011 and P = 0.041) compared to controls. No significant differences in cartilage, meniscus, and overall WORMS were found between the groups (P > 0.05). CONCLUSION: In comparison to nondiabetic controls, cartilage in DM subjects showed higher and more heterogeneous cartilage T2 values, indicating increased articular cartilage degeneration. This affected even more compartments in subjects with severe DM. LEVEL OF EVIDENCE: 2 Technical Efficacy: 5 J. Magn. Reson. Imaging 2018;47:380-390.

Conservatively treated knee injury is associated with knee cartilage matrix degeneration measured with MRI-based T2 relaxation times: data from the osteoarthritis initiative.

OBJECTIVE: To investigate the association of cartilage degeneration with previous knee injuries not undergoing surgery, determined by morphologic and quantitative 3-T magnetic resonance imaging (MRI). MATERIALS AND METHODS: We performed a nested cross-sectional study of right knee MRIs from participants in the Osteoarthritis Initiative (OAI) aged 45-79 with baseline Kellgren-Lawrence score of 0-2. Cases were 142 right knees of patients with self-reported history of injury limiting the ability to walk for at least 2 days. Controls were 426 right knees without history of injury, frequency-matched to cases on age, BMI, gender, KL scores and race (1:3 ratio). Cases and controls were compared using covariate-adjusted linear regression analysis, with the outcomes of region-specific T2 mean, laminar analysis and heterogeneity measured by texture analysis to investigate early cartilage matrix abnormalities and the Whole-Organ Magnetic Resonance Imaging Score (WORMS) to investigate morphologic knee lesions. RESULTS: Compared to control subjects, we found significantly higher mean T2 values in the injury [lateral tibia (28.10 ms vs. 29.11 ms, p = 0.001), medial tibia (29.70 ms vs. 30.40 ms, p = 0.014) and global knee cartilage (32.73 ms vs. 33.29 ms, p = 0.005)]. Injury subjects also had more heterogeneous cartilage as measured by GLCM texture contrast, variance and entropy (p < 0.05 in 14 out of 18 texture parameters). WORMS gradings were not significantly different between the two groups (p > 0.05). CONCLUSION: A history of knee injury not treated surgically is associated with higher and more heterogeneous T2 values, but not with morphologic knee abnormalities. Our findings suggest that significant, conservatively treated knee injuries are associated with permanent cartilage matrix abnormalities.