Combination of Radiological and Clinical Baseline Data for Outcome Prediction of Patients With an Acute Ischemic Stroke.

Background: Accurate prediction of clinical outcome is of utmost importance for choices regarding the endovascular treatment (EVT) of acute stroke. Recent studies on the prediction modeling for stroke focused mostly on clinical characteristics and radiological scores available at baseline. Radiological images are composed of millions of voxels, and a lot of information can be lost when representing this information by a single value. Therefore, in this study we aimed at developing prediction models that take into account the whole imaging data combined with clinical data available at baseline. Methods: We included 3,279 patients from the MR CLEAN Registry; a prospective, observational, multicenter registry of patients with ischemic stroke treated with EVT. We developed two approaches to combine the imaging data with the clinical data. The first approach was based on radiomics features, extracted from 70 atlas regions combined with the clinical data to train machine learning models. For the second approach, we trained 3D deep learning models using the whole images and the clinical data. Models trained with the clinical data only were compared with models trained with the combination of clinical and image data. Finally, we explored feature importance plots for the best models and identified many known variables and image features/brain regions that were relevant in the model decision process. Results: From 3,279 patients included, 1,241 (37%) patients had a good functional outcome [modified Rankin Scale (mRS) ≤ 2] and 1,954 (60%) patients had good reperfusion [modified Thrombolysis in Cerebral Infarction (eTICI) ≥ 2b]. There was no significant improvement by combining the image data to the clinical data for mRS prediction [mean area under the receiver operating characteristic (ROC) curve (AUC) of 0.81 vs. 0.80] above using the clinical data only, regardless of the approach used. Regarding predicting reperfusion, there was a significant improvement when image and clinical features were combined (mean AUC of 0.54 vs. 0.61), with the highest AUC obtained by the deep learning approach. Conclusions: The combination of radiomics and deep learning image features with clinical data significantly improved the prediction of good reperfusion. The visualization of prediction feature importance showed both known and novel clinical and imaging features with predictive values.

Statistical evaluation of diet-microbe associations.

BACKGROUND: Statistical evaluation of the association between microbial abundance and dietary variables can be done in various ways. Currently, there is no consensus on which methods are to be preferred in which circumstances. Application of particular methods seems to be based on the tradition of a particular research group, availability of experience with particular software, or depending on the outcomes of the analysis. RESULTS: We applied four popular methods including edgeR, limma, metagenomeSeq and shotgunFunctionalizeR, to evaluate the association between dietary variables and abundance of microbes. We found large difference in results between the methods. Our simulation studies revealed that no single method was optimal. CONCLUSIONS: We advise researchers to run multiple analyses and focus on the significant findings identified by multiple methods in order to achieve a better control of false discovery rate, although the false discovery rate can still be substantial.

Atherosclerosis in patients with cyanotic congenital heart disease.

BACKGROUND: Cyanotic patients with congenital heart disease (CHD) might be protected against atherosclerosis. METHODS AND RESULTS: Atherosclerotic risk factors and carotid intima - media thickness (IMT) were investigated in adults with cyanotic CHD and in unaffected age- and sex-matched controls. Fifty-four cyanotic patients (30 men, mean age 38, range 19-60 years) and 54 controls were included. Mean transcutaneous saturation of the cyanotic patients was 81+/-6%. Mean carotid IMT adjusted for age was significantly decreased in cyanotic patients compared to controls (0.55+/-0.1 mm vs 0.58+/-0.08 mm: DeltaIMT =0.04 mm [SE 0.015], P=0.01). In cyanotic patients lower total cholesterol levels were observed (4.4+/-1 mmol/L vs 4.9+/-1 mmol/L; P=0.02), as well as lower thrombocyte levels (173+/-81 x 10(9) /L vs 255+/-54 x 10(9) /L; P<0.01), higher bilirubin levels (18.6+/-11 micromol/L vs 12.7+/-6 micromol/L; P<0.01), and lower diastolic and systolic blood pressure (71+/-9 mmHg vs 76+/-9 mmHg, P<0.01; 113+/-14 mmHg vs 124+/-12 mmHg, P<0.01, respectively). CONCLUSIONS: In patients with cyanotic CHD carotid IMT, and hence atherosclerosis disease risk, was decreased. This might be due to a combination of reduced atherosclerotic risk factors such as lower blood pressure, lower total cholesterol levels, higher bilirubin levels and lower thrombocyte levels.

Long-term follow-up of homograft function after pulmonary valve replacement in patients with tetralogy of Fallot.

AIMS: To analyse the long-term outcomes after pulmonary valve replacement (PVR) in patients with a previous correction for tetralogy of Fallot. METHODS AND RESULTS: In a retrospective study, 158 adult patients with a diagnosis of tetralogy of Fallot, who had undergone a PVR after initial total correction in childhood, were identified from the CONCOR (CONgenital CORvitia) registry. All patients underwent 175 PVRs between June 1986 and June 2005. To analyse the predictors for homograft dysfunction and adverse events (death, reoperations, balloon angioplasty), Cox-regression analysis was performed. Overall freedom from significant homograft dysfunction was 66% after 5 years and 47% after 10 years. We could not identify predictors for combined homograft dysfunction. Event-free survival was 78% at 10 years and 68% at 15 years after PVR. Both early significant pulmonary regurgitation (PR) (HR 6.8, P = 0.017) and pulmonary stenosis (PS) (HR 4.0, P = 0.037) after surgery were associated with adverse events. When analysing direct post-operative PR or PS, we observed that in patients with severe, pre-operative PR, right ventricular aneurysm/patch resection resulted in a lower post-operative PR (mean difference grade 0.38 +/- 0.14, P = 0.01). Less significant post-operative PS was associated with a higher diameter of the homograft (HR 0.37, P = 0.006). CONCLUSION: While 47% of the patients in our study were free from homograft dysfunction at 10 years after PVR, event-free survival after PVR remained fairly good (78%). Significant residual lesions directly after surgery influenced event-free survival. A smaller diameter of the pulmonary homograft and severe pre-surgical PR were related to early homograft dysfunction after surgery.

Aortic root growth in men and women with the Marfan's syndrome.

The leading cause of premature death in patients with Marfan's syndrome (MS) is type A aortic dissection or rupture due to progressive aortic root dilation. The aim of this study was to analyze aortic root growth in 113 men and 108 women with MS. All patients were prospectively followed with serial echocardiograms of the native aortic root. At baseline, women had on average a 5-mm smaller aortic root diameter adjusted for age than men. Average aortic root growth was 0.42 mm/year (SE 0.05) in men and 0.38 mm/year (SE 0.04) in women. On the basis of aortic root growth rates, the men and women could be divided into 2 normally distributed subgroups: fast and slow growers. Approximately 1 in 7 men (1.5 mm/year, SE 0.5) and approximately 1 in 9 women (1.8 mm/year, SE 0.3) had fast-growing aortic root diameters. Significantly more type A dissections (25% vs 4%, p <0.001) were observed in fast growers than in slow growers; this was found in men and women. Type A dissections were observed in 4 men and 9 women. By reducing the cut-off value by 5 mm for elective aortic root replacement in women, type A dissections could have been prevented in 3 women. In conclusion, guidelines should take gender differences into account, and therefore, the investigators propose reducing the threshold for elective aortic root replacement in women with MS by 5 mm.

Outcome of pregnancy in patients after repair of aortic coarctation.

AIMS: Nowadays, most women born with aortic coarctation reach childbearing age. However, data on outcome of pregnancy in women after repair of aortic coarctation are scarce. The aim of this study was to report on maternal and neonatal outcome of pregnancy in women after aortic coarctation repair. METHODS AND RESULTS: The CONCOR national registry on congenital heart disease in The Netherlands was reviewed for women of childbearing age (> or =18 years old) with a history of aortic coarctation repair. Medical history and maternal, obstetrical, and neonatal outcome were determined. Fifty-four of the 100 women included had a history of pregnancy. The 54 women had 126 pregnancies resulting in 98 successful pregnancies, 22 miscarriages, and six abortions. The success rate was estimated as 0.778 (SE 0.002) including abortions and 0.817 (SE 0.002) excluding abortions. There were 85 vaginal deliveries, seven vaginal deliveries with epidural analgesia, and six caesarean sections. There were two neonatal deaths. A total of 26 pregnancies were complicated by a hypertensive disorder of pregnancy. There were 21 pregnancies in 14 women complicated by hypertension and five pregnancies in four women complicated by pre-eclampsia. The hypertension- and pre-eclampsia-probabilities were estimated as 0.183 (SE 0.285) and 0.061 (SE 0.211), respectively. During pregnancy, five patients had an increase > or =15 mmHg across the site of repair at echocardiography, but only one patient required reintervention for recoarctation after delivery. Four of the 98 children (4%) had a congenital heart defect. CONCLUSION: Pregnancy is well tolerated in women after repair of aortic coarctation. However, an excess of miscarriages and hypertensive disorders of pregnancy were found.

Pregnancy and aortic root growth in the Marfan syndrome: a prospective study.

AIMS: In women with Marfan syndrome pregnancy presents an increased risk of dilatation, dissection, and rupture of the aorta. The aim of this study was to investigate the influence of pregnancy on growth of the aortic root. METHODS AND RESULTS: Between 1993 and 2004 127 women with Marfan syndrome were prospectively followed; 61 women had one or more children; in 23 women, 33 pregnancies could be followed prospectively for aortic dimensions. Only one woman had suffered an aortic complication, a type A dissection (limited to the ascending aorta), before pregnancy. Out of 66 childless women a comparison group of 22 women was selected and individually matched. Mean initial aortic root diameter just before pregnancy was 37+/-5 mm (range 25-45). Before, during, and after pregnancy the overall individual aortic root diameter change (in 31 pregnancies) was not significant (P=0.77). Only the woman with a previous type A dissection developed an aortic complication (type B dissection) during her second pregnancy. No cardiac complications occurred in the other 22 women during their pregnancies. During a median follow-up of 6.4 years, no significant difference in growth of the aortic root was observed between the pregnancy group and the matched childless group (0.28 vs. 0.19 mm/year, P=0.08, respectively). CONCLUSION: Pregnancy in women with Marfan syndrome seems to be relatively safe up to an aortic root diameter of 45 mm, at least as far as our observed diameter range of 25-45 mm is concerned.

Predictive value of mild, residual descending aortic narrowing for blood pressure and vascular damage in patients after repair of aortic coarctation.

AIMS: The significance of mild residual descending aortic narrowing in post-coarctectomy patients is not known. The aim of our study was to investigate the influence of mild residual descending aortic narrowing on blood pressure and vascular damage in patients after repair of aortic coarctation. METHODS AND RESULTS: In 107 consecutive post-coarctectomy patients, magnetic resonance imaging, ambulatory blood pressure monitoring, and B-mode ultrasound of the carotid arteries were performed. A significant residual aortic narrowing was defined as: (i) a resting blood pressure gradient > or =30 mmHg with hypertension or exercise-induced hypertension (European Society of Cardiology guidelines); and/or (ii) a site of repair/diaphragmatic aorta ratio <0.7. Thirty-four patients (32%) had a significant residual aortic narrowing and were excluded from the analysis. Of the remaining 73 patients (43 male) with no or only mild residual descending aortic narrowing, median age was 29.8 years (range 17.1-52.5 years), mean age at repair 8.1 years (range 0.02-37.3 years), mean arm/leg gradient 2+/-12 mmHg, and mean common carotid intima-media thickness 0.612+/-0.118 mm. Thirty-three (45%) of these patients had hypertension. In multivariable regression analysis the site of repair/diaphragmatic aorta ratio was a strong and independent predictor of mean daytime systolic blood pressure (P<0.001) and common carotid intima-media thickness (P=0.027). CONCLUSION: Mild residual descending aortic narrowing in post-coarctectomy patients is independently associated with mean daytime blood pressure and carotid intima-media thickness. Our data suggest that a threshold for re-intervention of residual aortic narrowing lower than posed in current guidelines may be desirable to improve long-term outcome in these patients. However, further research on such aggressive interventional approaches is needed.