Deep Learning Staging of Liver Iron Content From Multiecho MR Images.

BACKGROUND: MRI represents the most established liver iron content (LIC) evaluation approach by estimation of liver T2* value, but it is dependent on the choice of the measurement region and the software used for image analysis. PURPOSE: To develop a deep-learning method for unsupervised classification of LIC from magnitude T2* multiecho MR images. STUDY TYPE: Retrospective. POPULATION/SUBJECTS: A total of 1069 thalassemia major patients enrolled in the core laboratory of the Myocardial Iron Overload in Thalassemia (MIOT) network, which were included in the training (80%) and test (20%) sets. Twenty patients from different MRI vendors included in the external test set. FIELD STRENGTH/SEQUENCE: A5 T, T2* multiecho magnitude images. ASSESSMENT: Four deep-learning convolutional neural networks (HippoNet-2D, HippoNet-3D, HippoNet-LSTM, and an ensemble network HippoNet-Ensemble) were used to achieve unsupervised staging of LIC using five classes (normal, borderline, middle, moderate, severe). The training set was employed to construct the deep-learning model. The performance of the LIC staging model was evaluated in the test set and in the external test set. The model's performances were assessed by evaluating the accuracy, sensitivity, and specificity with respect to the ground truth labels obtained by T2* measurements and by comparison with operator-induced variability originating from different region of interest (ROI) placements. STATISTICAL TESTS: The network's performances were evaluated by single-class accuracy, specificity, and sensitivity and compared by one-way repeated measures analysis of variance (ANOVA) and one-way ANOVA. RESULTS: HippoNet-Ensemble reached an accuracy significantly higher than the other networks, and a sensitivity and specificity higher than HippoNet-LSTM. Accuracy, sensitivity, and specificity values for the LIC stages were: normal: 0.96/0.93/0.97, borderline: 0.95/0.85/0.98, mild: 0.96/0.88/0.98, moderate: 0.95/0.89/0.97, severe: 0.97/0.95/0.98. Correctly staging of cases was in the range of 85%-95%, depending on the LIC class. Multiclass accuracy was 0.90 against 0.92 for the interobserver variability. DATA CONCLUSION: The proposed HippoNet-Ensemble network can perform unsupervised LIC staging and achieves good prognostic performance. EVIDENCE LEVEL: 4 TECHNICAL EFFICACY: Stage 2.

Increased myocardial extracellular volume is associated with myocardial iron overload and heart failure in thalassemia major.

OBJECTIVES: Myocardial extracellular volume (ECV) by cardiovascular magnetic resonance (CMR) is a surrogate marker of diffuse fibrosis. We evaluated the association between ECV and demographics, CMR findings, and cardiac involvement in patients with thalassemia major (TM). METHODS: A total of 108 ß-TM patients (62 females, 40.16 ± 8.83 years), consecutively enrolled in the Extension-Myocardial Iron Overload in Thalassemia Network, and 16 healthy subjects (6 females, 37.12 ± 16.13 years) underwent CMR. The protocol included assessment of T2*, native T1, and T2 values in all 16 myocardial segments for myocardial iron overload (MIO) quantification, cine images for left ventricular (LV) function quantification, post-contrast T1 mapping for ECV calculation, and late gadolinium enhancement (LGE) technique for replacement myocardial fibrosis detection. RESULTS: Global ECV values were significantly higher in females than in males. Global ECV values were significantly higher in patients with significant MIO (global heart T2* < 20 ms) than in patients without significant MIO, and both groups exhibited higher global ECV values than healthy subjects. No association was detected between native T1 and ECV values, while patients with reduced global heart T2 values showed significantly higher global ECV values than patients with normal and increased global heart T2. Global ECV values were not correlated with LV function/size and were comparable between patients with and without LGE. Compared to patients without heart failure, patients with a history of heart failure (N = 10) showed significantly higher global heart ECV values. CONCLUSION: In TM, increased myocardial ECV, potentially reflecting diffuse interstitial fibrosis, is associated with MIO and heart failure. KEY POINTS: • CMR-derived myocardial extracellular volume is increased in thalassemia major patients, irrespective of the presence of late gadolinium enhancement. • In thalassemia major, myocardial iron overload contributes to the increase in myocardial ECV, which potentially reflects diffuse interstitial fibrosis and is significantly associated with a history of heart failure.

Pancreatic fatty replacement as risk marker for altered glucose metabolism and cardiac iron and complications in thalassemia major.

OBJECTIVES: This multicenter study assessed the extent of pancreatic fatty replacement and its correlation with demographics, iron overload, glucose metabolism, and cardiac complications in a cohort of well-treated patients with thalassemia major (TM). METHODS: We considered 308 TM patients (median age: 39.79 years; 182 females) consecutively enrolled in the Extension-Myocardial Iron Overload in Thalassemia Network. Magnetic resonance imaging was used to quantify iron overload (IO) and pancreatic fat fraction (FF) by T2* technique, cardiac function by cine images, and to detect replacement myocardial fibrosis by late gadolinium enhancement technique. The glucose metabolism was assessed by the oral glucose tolerance test. RESULTS: Pancreatic FF was associated with age, body mass index, and history of hepatitis C virus infection. Patients with normal glucose metabolism showed a significantly lower pancreatic FF than patients with impaired fasting glucose (p = 0.030), impaired glucose tolerance (p < 0.0001), and diabetes (p < 0.0001). A normal pancreatic FF (< 6.6%) showed a negative predictive value of 100% for abnormal glucose metabolism. A pancreatic FF > 15.33% predicted the presence of abnormal glucose metabolism. Pancreas FF was inversely correlated with global pancreas and heart T2* values. A normal pancreatic FF showed a negative predictive value of 100% for cardiac iron. Pancreatic FF was significantly higher in patients with myocardial fibrosis (p = 0.002). All patients with cardiac complications had fatty replacement, and they showed a significantly higher pancreatic FF than complications-free patients (p = 0.002). CONCLUSION: Pancreatic FF is a risk marker not only for alterations of glucose metabolism, but also for cardiac iron and complications, further supporting the close link between pancreatic and cardiac disease. KEY POINTS: • In thalassemia major, pancreatic fatty replacement by MRI is a frequent clinical entity, predicted by a pancreas T2* < 20.81 ms and associated with a higher risk of alterations in glucose metabolism. • In thalassemia major, pancreatic fatty replacement is a strong risk marker for cardiac iron, replacement fibrosis, and complications, highlighting a deep connection between pancreatic and cardiac impairment.

Myocardial iron overload by cardiovascular magnetic resonance native segmental T1 mapping: a sensitive approach that correlates with cardiac complications.

BACKGROUND: We compared cardiovascular magnetic resonance segmental native T1 against T2* values for the detection of myocardial iron overload (MIO) in thalassaemia major and we evaluated the clinical correlates of native T1 measurements. METHODS: We considered 146 patients (87 females, 38.7 ± 11.1 years) consecutively enrolled in the Extension-Myocardial Iron Overload in Thalassaemia Network. T1 and T2* values were obtained in the 16 left ventricular (LV) segments. LV function parameters were quantified by cine images. Post-contrast late gadolinium enhancement (LGE) and T1 images were acquired. RESULTS: 64.1% of segments had normal T2* and T1 values while 10.1% had pathologic T2* and T1 values. In 526 (23.0%) segments, there was a pathologic T1 and a normal T2* value while 65 (2.8%) segments had a pathologic T2* value but a normal T1 and an extracellular volume (ECV) ≥ 25% was detected in 16 of 19 segments where ECV was quantified. Global native T1 was independent from gender or LV function but decreased with increasing age. Patients with replacement myocardial fibrosis had significantly lower native global T1. Patients with cardiac complications had significantly lower native global T1. CONCLUSIONS: The combined use of both segmental native T1 and T2* values could improve the sensitivity for detecting MIO. Native T1 is associated with cardiac complications in thalassaemia major.

Investigating multiple sclerosis genetic susceptibility on the founder population of east-central Sardinia via association and linkage analysis of immune-related loci.

BACKGROUND: A wealth of single-nucleotide polymorphisms (SNPs) responsible for multiple sclerosis (MS) susceptibility have been identified; however, they explain only a fraction of MS heritability. OBJECTIVES: We contributed to discovery of new MS susceptibility SNPs by studying a founder population with high MS prevalence. METHODS: We analyzed ImmunoChip data from 15 multiplex families and 94 unrelated controls from the Nuoro Province, Sardinia, Italy. We tested each SNP for both association and linkage with MS, the linkage being explored in terms of identity-by-descent (IBD) sharing excess and using gene dropping to compute a corresponding empirical p-value. By targeting regions that are both associated and in linkage with MS, we increase chances of identifying interesting genomic regions. RESULTS: We identified 486 MS-associated (p < 1 × 10-4) and 18,426 MS-linked (p < 0.05) SNPs. A total of 111 loci were both linked and associated with MS, 18 of them pointing to 14 non-major histocompatibility complex (MHC) genes, and 93 of them located in the MHC region. CONCLUSION: We discovered new suggestive signals and confirmed some previously identified ones. We believe this to represent a significant step toward an understanding of the genetic basis of MS.

Myocardial tissue characterization by segmental T2 mapping in thalassaemia major: detecting inflammation beyond iron.

AIMS: We measured myocardial T2 values by a segmental approach in thalassaemia major (TM) patients, comparing such values against T2* values for the detection of myocardial iron overload (MIO), evaluating their potential in detecting subclinical inflammation, and correlating with clinical status. METHODS AND RESULTS: One-hundred and sixty-six patients (102 females, 38.29 ± 11.49years) enrolled in the Extension-Myocardial Iron Overload in Thalassemia Network underwent magnetic resonance imaging for the assessment of hepatic, pancreatic, and cardiac iron overload (T2* technique), of biventricular function (cine images), and of replacement myocardial fibrosis [late gadolinium enhancement (LGE)]. T2 and T2* values were quantified in all 16 myocardial segments, and the global value was the mean of all segments. Global heart T2 values were significantly higher in TM than in a cohort of 80 healthy subjects. T2 and T2* values were significantly correlated. Out of the 25 patients with a decreased global heart T2* value, 11 (44.0%) had reduced T2 values. No patient with a normal T2* value had a decreased T2 value.Eleven (6.6%) patients had a decreased global heart T2 value, 74 (44.6%) a normal global heart T2 value, and 81 (48.8%) an increased global heart T2 value. Biventricular function was comparable amongst the three groups, whilst LGE was significantly more frequent in patients with reduced vs. increased global heart T2 value. Compared with the other two groups, patients with reduced T2 values had significantly higher hepatic and pancreatic iron deposition. CONCLUSION: In TM, T2 mapping does not offer any advantage in terms of sensitivity for MIO assessment but detects subclinical myocardial inflammation.

Feasibility, reproducibility, and reliability for the T*2 iron evaluation at 3 T in comparison with 1.5 T.

This study aimed to determine the feasibility, reproducibility, and reliability of the multiecho T*(2) Magnetic resonance imaging technique at 3 T for myocardial and liver iron burden quantification and the relationship between T*(2) values at 3 and 1.5 T. Thirty-eight transfusion-dependent patients and 20 healthy subjects were studied. Cardiac segmental and global T*(2) values were calculated after developing a correction map to compensate the artifactual T*(2) variations. The hepatic T*(2) value was determined over a region of interest. The intraoperator and interoperator reproducibility for T*(2) measurements at 3 T was good. A linear relationship was found between patients' R *2 (1000/T*(2) ) values at 3 and 1.5 T. Segmental correction factors were significantly higher at 3 T. A conversion formula returning T*(2) values at 1.5 T from values at 3 T was proposed. A good diagnostic reliability for T*(2) assessment at 3 T was demonstrated. Lower limits of normal for 3 T T*(2) values were 23.3 ms, 21.1 ms, and 11.7 ms, for the global heart, mid-ventricular septum, and liver, respectively. In conclusion, T*(2) quantification of iron burden in the mid-ventricular septum, global heart, and no heavy-moderate livers resulted to be feasible, reproducible, and reliable at 3 T. Segmental heart T*(2) analysis at 3 T may be challenging due to significantly higher susceptibility artifacts.

Gender Differences in Knowledge and Perception of Cardiovascular Disease among Italian Thalassemia Major Patients.

We evaluated gender differences in knowledge and perception of cardiovascular disease (CVD) among Italian thalassemia major (TM) patients. An anonymous questionnaire was completed by 139 ß-TM patients (87 (62.7%) females, 40.90 ± 8.03 years). Compared to females, males showed a significantly higher frequency of CVDs, and they less frequently selected tumors in general as the greatest health problem for people of the same age and gender (48.1% vs. 66.7%; p = 0.031) and as the greatest danger to their future health (26.9% vs. 43.7%; p = 0.048). CVDs were designated as the greatest danger to their future health by a significantly higher percentage of males than females (53.8% vs. 36.8%; p = 0.048). Both males and females showed a good knowledge of cardiovascular risk factors and preventive measures for CVDs. No gender differences were detected in the subjective well-being and the perceived cardiovascular risk. The perceived risk was not influenced by age, presence of cardiovascular risk factors, or disease, but no patient with a low perceived CVD risk had myocardial iron overload. Our findings highlight the need to implement future educational programs aimed at increasing the awareness of CVD as the greatest health issue, especially among the female TM population, and at informing TM patients of the different actors, besides iron, that play a role in the development of cardiovascular complications.

Fast generation of T2* maps in the entire range of clinical interest: application to thalassemia major patients.

T2* maps obtained by the processing of multiecho MR sequences can be useful in several clinical applications. T2* map generation procedures should join a processing time compatible with on-line image analysis with a good precision in the entire T2* range of clinical interest. Fast generation of T2* maps can be achieved by the estimation of the T2* values by the weighted linear fitting of the logarithm of the signal (WLSL) method. This approach fails if the signal decay diverges from a pure exponential decay, as happens at low T2* values where the rapid decay in the signal intensity leads to a plateau in the later echo times (TE). The proposed method implements the automatic truncation of the signal decay curves to be fitted in order to compensate for the signal collapse at low T2* values, allowing the extension of the WLSL method through the entire clinical range of T2* values. Validation was performed on synthetic images and on 60 thalassemia major patients with different levels of myocardial iron overload. Phantom experiments showed that a 5% fitting error threshold represented the best compromise between T2* value measurement precision and processing time. A good agreement was found between T2* map pixel-wise measurements and ROI-based measurements performed by expert readers (CoV=1.84% in global heart T2*, CoV=5.8% in segmental analysis). In conclusion, the developed procedure was effective in generating correct T2* maps for the entire T2* clinical range.

Association between protective and deleterious HLA alleles with multiple sclerosis in Central East Sardinia.

The human leukocyte antigen (HLA) complex on chromosome 6p21 has been unambiguously associated with multiple sclerosis (MS). The complex features of the HLA region, especially its high genic content, extreme polymorphism, and extensive linkage disequilibrium, has prevented to resolve the nature of HLA association in MS. We performed a family based association study on the isolated population of the Nuoro province (Sardinia) to clarify the role of HLA genes in MS. The main stage of our study involved an analysis of the ancestral haplotypes A2Cw7B58DR2DQ1 and A30Cw5B18DR3DQ2. On the basis of a multiplicative model, the effect of the first haplotype is protective with an odds ratio (OR) = 0.27 (95% confidence interval CI 0.13-0.57), while that of the second is deleterious, OR 1.78 (95% CI 1.26-2.50). We found both class I (A, Cw, B) and class II (DR, DQ) loci to have an effect on MS susceptibility, but we saw that they act independently from each other. We also performed an exploratory analysis on a set of 796 SNPs in the same HLA region. Our study supports the claim that Class I and Class II loci act independently on MS susceptibility and this has a biological explanation. Also, the analysis of SNPs suggests that there are other HLA genes involved in MS, but replication is needed. This opens up new perspective on the study of MS.

Association between the ACCN1 gene and multiple sclerosis in Central East Sardinia.

Multiple genome screens have been performed to identify regions in linkage or association with Multiple Sclerosis (MS, OMIM 126200), but little overlap has been found among them. This may be, in part, due to a low statistical power to detect small genetic effects and to genetic heterogeneity within and among the studied populations. Motivated by these considerations, we studied a very special population, namely that of Nuoro, Sardinia, Italy. This is an isolated, old, and genetically homogeneous population with high prevalence of MS. Our study sample includes both nuclear families and unrelated cases and controls. A multi-stage study design was adopted. In the first stage, microsatellites were typed in the 17q11.2 region, previously independently found to be in linkage with MS. One significant association was found at microsatellite D17S798. Next, a bioinformatic screening of the region surrounding this marker highlighted an interesting candidate MS susceptibility gene: the Amiloride-sensitive Cation Channel Neuronal 1 (ACCN1) gene. In the second stage of the study, we resequenced the exons and the 3' untranslated (UTR) region of ACCN1, and investigated the MS association of Single Nucleotide Polymorphisms (SNPs) identified in that region. For this purpose, we developed a method of analysis where complete, phase-solved, posterior-weighted haplotype assignments are imputed for each study individual from incomplete, multi-locus, genotyping data. The imputed assignments provide an input to a number of proposed procedures for testing association at a microsatellite level or of a sequence of SNPs. These include a Mantel-Haenszel type test based on expected frequencies of pseudocase/pseudocontrol haplotypes, as well as permutation based tests, including a combination of permutation and weighted logistic regression analysis. Application of these methods allowed us to find a significant association between MS and the SNP rs28936 located in the 3' UTR segment of ACCN1 with p = 0.0004 (p = 0.002, after adjusting for multiple testing). This result is in tune with several recent experimental findings which suggest that ACCN1 may play an important role in the pathogenesis of MS.

Two distinctive HLA haplotypes harbor the B27 alleles negatively or positively associated with ankylosing spondylitis in Sardinia: implications for disease pathogenesis.

OBJECTIVE: To compare haplotype distribution in HLA-B27-positive patients with ankylosing spondylitis (AS) and healthy control subjects possessing either AS-associated HLA-B27 alleles or the non-AS-associated HLA-B*2709 allele. METHODS: DNA samples from 47 HLA-B27-positive patients with AS and 76 HLA-B27-positive healthy controls (19 positive and 57 negative for B*2709) living in different areas of Sardinia were collected and typed for HLA class I and class II alleles. The third exon of the B27 gene was analyzed for the presence of Asp(116) or His(116), which differentiates B*2709 from the other two B27 subtypes (B*2705 and B*2702) that are mostly found in Sardinia. The parents of 6 subjects positive for B*2709 were also typed for HLA class I and class II alleles. Statistical analysis was performed by Fisher's exact test. RESULTS: In Sardinia, the B27 alleles conferring susceptibility to AS appear to be more frequently carried by a haplotype (A2;B27;Cw2;DR16) that reaches its highest frequency in patients with AS (A2 80.8%, B27 100%, Cw2 83%, and DR16 74.5%). Conversely, the non-AS-associated B*2709 allele is more frequently found together with other HLA alleles whose frequencies are inversely correlated with the disease (A32 or A30, Cw1, and DR12). Familial analysis of 6 subjects positive for HLA-B*2709 confirmed the existence of a "Sardinian" haplotype that is not associated with AS (A32;B*2709;Cw1;DR12). CONCLUSION: In Sardinia, 2 distinct haplotypes harbor the non-AS-associated HLA-B*2709 allele or the AS-associated B27 alleles. Our findings are compatible with the hypothesis that other genes within the HLA region besides HLA-B27 may play some role in conferring susceptibility to AS.