A multi-tissue, splicing-based joint transcriptome-wide association study identifies susceptibility genes for breast cancer.

Splicing-based transcriptome-wide association studies (splicing-TWASs) of breast cancer have the potential to identify susceptibility genes. However, existing splicing-TWASs test the association of individual excised introns in breast tissue only and thus have limited power to detect susceptibility genes. In this study, we performed a multi-tissue joint splicing-TWAS that integrated splicing-TWAS signals of multiple excised introns in each gene across 11 tissues that are potentially relevant to breast cancer risk. We utilized summary statistics from a meta-analysis that combined genome-wide association study (GWAS) results of 424,650 women of European ancestry. Splicing-level prediction models were trained in GTEx (v.8) data. We identified 240 genes by the multi-tissue joint splicing-TWAS at the Bonferroni-corrected significance level; in the tissue-specific splicing-TWAS that combined TWAS signals of excised introns in genes in breast tissue only, we identified nine additional significant genes. Of these 249 genes, 88 genes in 62 loci have not been reported by previous TWASs, and 17 genes in seven loci are at least 1 Mb away from published GWAS index variants. By comparing the results of our splicing-TWASs with previous gene-expression-based TWASs that used the same summary statistics and expression prediction models trained in the same reference panel, we found that 110 genes in 70 loci that are identified only by the splicing-TWASs. Our results showed that for many genes, expression quantitative trait loci (eQTL) did not show a significant impact on breast cancer risk, whereas splicing quantitative trait loci (sQTL) showed a strong impact through intron excision events.

Expression- and splicing-based multi-tissue transcriptome-wide association studies identified multiple genes for breast cancer by estrogen-receptor status.

BACKGROUND: Although several transcriptome-wide association studies (TWASs) have been performed to identify genes associated with overall breast cancer (BC) risk, only a few TWAS have explored the differences in estrogen receptor-positive (ER+) and estrogen receptor-negative (ER-) breast cancer. Additionally, these studies were based on gene expression prediction models trained primarily in breast tissue, and they did not account for alternative splicing of genes. METHODS: In this study, we utilized two approaches to perform multi-tissue TWASs of breast cancer by ER subtype: (1) an expression-based TWAS that combined TWAS signals for each gene across multiple tissues and (2) a splicing-based TWAS that combined TWAS signals of all excised introns for each gene across tissues. To perform this TWAS, we utilized summary statistics for ER + BC from the Breast Cancer Association Consortium (BCAC) and for ER- BC from a meta-analysis of BCAC and the Consortium of Investigators of Modifiers of BRCA1 and BRCA2 (CIMBA). RESULTS: In total, we identified 230 genes in 86 loci that were associated with ER + BC and 66 genes in 29 loci that were associated with ER- BC at a Bonferroni threshold of significance. Of these genes, 2 genes associated with ER + BC at the 1q21.1 locus were located at least 1 Mb from published GWAS hits. For several well-studied tumor suppressor genes such as TP53 and CHEK2 which have historically been thought to impact BC risk through rare, penetrant mutations, we discovered that common variants, which modulate gene expression, may additionally contribute to ER + or ER- etiology. CONCLUSIONS: Our study comprehensively examined how differences in common variation contribute to molecular differences between ER + and ER- BC and introduces a novel, splicing-based framework that can be used in future TWAS studies.

Gene-based association tests in family samples using GWAS summary statistics.

Genome-wide association studies (GWAS) have led to rapid growth in detecting genetic variants associated with various phenotypes. Owing to a great number of publicly accessible GWAS summary statistics, and the difficulty in obtaining individual-level genotype data, many existing gene-based association tests have been adapted to require only GWAS summary statistics rather than individual-level data. However, these association tests are restricted to unrelated individuals and thus do not apply to family samples directly. Moreover, due to its flexibility and effectiveness, the linear mixed model has been increasingly utilized in GWAS to handle correlated data, such as family samples. However, it remains unknown how to perform gene-based association tests in family samples using the GWAS summary statistics estimated from the linear mixed model. In this study, we show that, when family size is negligible compared to the total sample size, the diagonal block structure of the kinship matrix makes it possible to approximate the correlation matrix of marginal Z scores by linkage disequilibrium matrix. Based on this result, current methods utilizing summary statistics for unrelated individuals can be directly applied to family data without any modifications. Our simulation results demonstrate that this proposed strategy controls the type 1 error rate well in various situations. Finally, we exemplify the usefulness of the proposed approach with a dental caries GWAS data set.

Multi-tissue transcriptome-wide association studies identified 235 genes for intrinsic subtypes of breast cancer.

BACKGROUND: Although genome-wide association studies (GWAS) of breast cancer (BC) identified common variants which differ between intrinsic subtypes, genes through which these variants act to impact BC risk have not been fully established. Transcriptome-wide association studies (TWAS) have identified genes associated with overall BC risk, but subtype-specific differences are largely unknown. METHODS: We performed two multi-tissue TWAS for each BC intrinsic subtype, including an expression-based approach that collated TWAS signals from expression quantitative trait loci (eQTLs) across multiple tissues and a novel splicing-based approach that collated signals from splicing QTLs (sQTLs) across intron clusters and subsequently across tissues. We used summary statistics for five intrinsic subtypes including Luminal A-like, Luminal B-like, Luminal B/HER2-negative-like, HER2-enriched-like, and triple-negative BC, generated from 106 278 BC cases and 91 477 controls in the Breast Cancer Association Consortium. RESULTS: Overall, we identified 235 genes in 88 loci that were associated with at least one of the five intrinsic subtypes. Most genes were subtype-specific, and many have not been reported in previous TWAS. We discovered common variants that modulate expression of CHEK2 confer increased risk to Luminal A-like BC, in contrast to the viewpoint that CHEK2 primarily harbors rare, penetrant mutations. Additionally, our splicing-based TWAS provided population-level support for MDM4 splice variants that increased the risk of triple-negative BC. CONCLUSION: Our comprehensive, multi-tissue TWAS corroborated previous GWAS loci for overall BC risk and intrinsic subtypes, while underscoring how common variation that impacts expression and splicing of genes in multiple tissue types can be used to further elucidate the etiology of BC.

Novel breast cancer susceptibility loci under linkage peaks identified in African ancestry consortia.

BACKGROUND: Expansion of genome-wide association studies across population groups is needed to improve our understanding of shared and unique genetic contributions to breast cancer. We performed association and replication studies guided by a priori linkage findings from African ancestry (AA) relative pairs. METHODS: We performed fixed-effect inverse-variance weighted meta-analysis under three significant AA breast cancer linkage peaks (3q26-27, 12q22-23, and 16q21-22) in 9241 AA cases and 10 193 AA controls. We examined associations with overall breast cancer as well as estrogen receptor (ER)-positive and negative subtypes (193,132 SNPs). We replicated associations in the African-ancestry Breast Cancer Genetic Consortium (AABCG). RESULTS: In AA women, we identified two associations on chr12q for overall breast cancer (rs1420647, OR = 1.15, p = 2.50×10-6; rs12322371, OR = 1.14, p = 3.15×10-6), and one for ER-negative breast cancer (rs77006600, OR = 1.67, p = 3.51×10-6). On chr3, we identified two associations with ER-negative disease (rs184090918, OR = 3.70, p = 1.23×10-5; rs76959804, OR = 3.57, p = 1.77×10-5) and on chr16q we identified an association with ER-negative disease (rs34147411, OR = 1.62, p = 8.82×10-6). In the replication study, the chr3 associations were significant and effect sizes were larger (rs184090918, OR: 6.66, 95% CI: 1.43, 31.01; rs76959804, OR: 5.24, 95% CI: 1.70, 16.16). CONCLUSION: The two chr3 SNPs are upstream to open chromatin ENSR00000710716, a regulatory feature that is actively regulated in mammary tissues, providing evidence that variants in this chr3 region may have a regulatory role in our target organ. Our study provides support for breast cancer variant discovery using prioritization based on linkage evidence.

A joint transcriptome-wide association study across multiple tissues identifies candidate breast cancer susceptibility genes.

Genome-wide association studies (GWASs) have identified more than 200 genomic loci for breast cancer risk, but specific causal genes in most of these loci have not been identified. In fact, transcriptome-wide association studies (TWASs) of breast cancer performed using gene expression prediction models trained in breast tissue have yet to clearly identify most target genes. To identify candidate genes, we performed a GWAS analysis in a breast cancer dataset from UK Biobank (UKB) and combined the results with the GWAS results of the Breast Cancer Association Consortium (BCAC) by a meta-analysis. Using the summary statistics from the meta-analysis, we performed a joint TWAS analysis that combined TWAS signals from multiple tissues. We used expression prediction models trained in 11 tissues that are potentially relevant to breast cancer from the Genotype-Tissue Expression (GTEx) data. In the GWAS analysis, we identified eight loci distinct from those reported previously. In the TWAS analysis, we identified 309 genes at 108 genomic loci to be significantly associated with breast cancer at the Bonferroni threshold. Of these, 17 genes were located in eight regions that were at least 1 Mb away from published GWAS hits. The remaining TWAS-significant genes were located in 100 known genomic loci from previous GWASs of breast cancer. We found that 21 genes located in known GWAS loci remained statistically significant after conditioning on previous GWAS index variants. Our study provides insights into breast cancer genetics through mapping candidate target genes in a large proportion of known GWAS loci and discovering multiple new loci.

Characterizing the distribution of alterations in mesothelioma and their correlation to morphology.

OBJECTIVES: Mesothelioma is a lethal disease that arises from the serosal lining of organ cavities. Several recurrent alterations have been observed in pleural and peritoneal -mesotheliomas, including in BAP1, NF2, and CDKN2A. Although specific histopathologic parameters have been correlated with prognosis, it is not as well known whether genetic alterations correlate with histologic findings. METHODS: We reviewed 131 mesotheliomas that had undergone next-generation sequencing (NGS) at our institutions after pathologic diagnosis. There were 109 epithelioid mesotheliomas, 18 biphasic mesotheliomas, and 4 sarcomatoid mesotheliomas. All our biphasic and sarcomatoid cases arose in the pleura. Of the epithelioid mesotheliomas, 73 were from the pleura and 36 were from the peritoneum. On average, patients were 66 years of age (range, 26-90 years) and predominantly male (92 men, 39 women). RESULTS: The most common alterations identified were in BAP1, CDKN2A, NF2, and TP53. Twelve mesotheliomas did not show a pathogenic alteration on NGS. For epithelioid mesotheliomas in the pleura, the presence of an alteration in BAP1 correlated with low nuclear grade (P = .04), but no correlation was found in the peritoneum (P = .62). Similarly, there was no correlation between the amount of solid architecture in epithelioid mesotheliomas and any alterations in the pleura (P = .55) or peritoneum (P = .13). For biphasic mesotheliomas, cases with either no alteration detected or with an alteration in BAP1 were more likely to be epithelioid predominant (>50% of the tumor, P = .0001), and biphasic mesotheliomas with other alterations detected and no alteration in BAP1 were more likely to be sarcomatoid predominant (>50% of the tumor, P = .0001). CONCLUSIONS: This study demonstrates a significant association between morphologic features associated with a better prognosis and an alteration in BAP1.

Outcomes of Universal Point-of-Care Genetic Testing in Diverse Patients With Pancreatic Ductal Adenocarcinoma.

PURPOSE: Guidelines recommend all patients with pancreatic ductal adenocarcinoma (PDAC) undergo germline genetic testing (GT). Rates of recommendation and completion of GT among diverse patients with PDAC are not known. The aim was to determine rates of recommendation and completion of point-of-care GT in diverse patients with PDAC. METHODS: A retrospective review of patients with PDAC seen at an academic center between April 2019 and December 2020 was performed. Recommendation, completion and results of point-of-care GT, and demographic and clinical factors were recorded. Univariate and multivariate analyses of GT were performed using the chi-square test and logistic regression. RESULTS: In total, 579 patients with PDAC were included. The median age at diagnosis was 67 years; 52% were male; 63% were non-Hispanic White (NHW) patients, and 20% were African American (AA) patients. GT was performed in 216 (37%) patients. Of those tested, 47 (22%) had a pathogenic/likely pathogenic variant identified of which 25 (12%) were in PDAC-associated genes. On multivariate analysis, age, NHW race, personal and family cancer history, medical oncology visit, and number of visits were independent predictors of GT completion. AA patients had significantly lower rates of recommendation and completion of GT compared with NHW patients. CONCLUSION: Point-of-care GT in patients with PDAC is unacceptably low, especially among AA patients. Testing disparity might be due to lack of provider recommendation more than patient uptake. Lack of testing leads to missed opportunities for potential targeted therapies, improved outcomes, and identification of at-risk family members who could potentially benefit from surveillance.

Allogeneic hematopoietic stem cell transplant outcomes in adults with inherited myeloid malignancies.

There is increasing recognition that pathogenic germ line variants drive the development of hematopoietic cancers in many individuals. Currently, patients with hereditary hematologic malignancies (HHMs) receive similar standard therapies and hematopoietic stem cell transplant (HSCT) approaches as those with sporadic disease. We hypothesize that patients with myeloid malignancies and deleterious germ line predisposition variants have different posttransplant outcomes than those without such alleles. We studied 472 patients with myeloid neoplasms, of whom 26% had deleterious germ line variants and 34% underwent HSCT. Deleterious germ line variants in CHEK2 and DDX41 were most commonly seen in American and Australian cohorts, respectively. Patients with deleterious germ line DDX41 variants had a higher incidence of severe (stage 3-4) acute graft-versus-host disease (GVHD) (38%) than recipients with deleterious CHEK2 variants (0%), other HHM variants (12%), or patients without such germ line variants (9%) (P = .002). Importantly, the use of posttransplant cyclophosphamide reduced the risk of severe acute GVHD in patients receiving HSCT for deleterious germ line DDX41-associated myeloid neoplasms (0% vs 53%, P = .03). Based on these results, we advocate the use of posttransplant cyclophosphamide when individuals with deleterious germ line DDX41 variants undergo allogeneic HSCT for myeloid malignancies, even when transplantation has been performed using wild-type donors.

MOVER-R and Penalized MOVER-R Confidence Intervals for the Ratio of Two Quantities.

Developing a confidence interval for the ratio of two quantities is an important task in statistics because of its omnipresence in real world applications. For such a problem, the MOVER-R (method of variance recovery for the ratio) technique, which is based on the recovery of variance estimates from confidence limits of the numerator and the denominator separately, was proposed as a useful and efficient approach. However, this method implicitly assumes that the confidence interval for the denominator never includes zero, which might be violated in practice. In this article, we first use a new framework to derive the MOVER-R confidence interval, which does not require the above assumption and covers the whole parameter space. We find that MOVER-R can produce an unbounded confidence interval, just like the well-known Fieller method. To overcome this issue, we further propose the penalized MOVER-R. We prove that the new method differs from MOVER-R only at the second order. It, however, always gives a bounded and analytic confidence interval. Through simulation studies and a real data application, we show that the penalized MOVER-R generally provides a better confidence interval than MOVER-R in terms of controlling the coverage probability and the median width.

Polygenic risk scores for prediction of breast cancer risk in women of African ancestry: a cross-ancestry approach.

Polygenic risk scores (PRSs) are useful for predicting breast cancer risk, but the prediction accuracy of existing PRSs in women of African ancestry (AA) remains relatively low. We aim to develop optimal PRSs for the prediction of overall and estrogen receptor (ER) subtype-specific breast cancer risk in AA women. The AA dataset comprised 9235 cases and 10 184 controls from four genome-wide association study (GWAS) consortia and a GWAS study in Ghana. We randomly divided samples into training and validation sets. We built PRSs using individual-level AA data by a forward stepwise logistic regression and then developed joint PRSs that combined (1) the PRSs built in the AA training dataset and (2) a 313-variant PRS previously developed in women of European ancestry. PRSs were evaluated in the AA validation set. For overall breast cancer, the odds ratio per standard deviation of the joint PRS in the validation set was 1.34 [95% confidence interval (CI): 1.27-1.42] with the area under receiver operating characteristic curve (AUC) of 0.581. Compared with women with average risk (40th-60th PRS percentile), women in the top decile of the PRS had a 1.98-fold increased risk (95% CI: 1.63-2.39). For PRSs of ER-positive and ER-negative breast cancer, the AUCs were 0.608 and 0.576, respectively. Compared with existing methods, the proposed joint PRSs can improve prediction of breast cancer risk in AA women.

Novel strategy for disease risk prediction incorporating predicted gene expression and DNA methylation data: a multi-phased study of prostate cancer.

BACKGROUND: DNA methylation and gene expression are known to play important roles in the etiology of human diseases such as prostate cancer (PCa). However, it has not yet been possible to incorporate information of DNA methylation and gene expression into polygenic risk scores (PRSs). Here, we aimed to develop and validate an improved PRS for PCa risk by incorporating genetically predicted gene expression and DNA methylation, and other genomic information using an integrative method. METHODS: Using data from the PRACTICAL consortium, we derived multiple sets of genetic scores, including those based on available single-nucleotide polymorphisms through widely used methods of pruning and thresholding, LDpred, LDpred-funt, AnnoPred, and EBPRS, as well as PRS constructed using the genetically predicted gene expression and DNA methylation through a revised pruning and thresholding strategy. In the tuning step, using the UK Biobank data (1458 prevalent cases and 1467 controls), we selected PRSs with the best performance. Using an independent set of data from the UK Biobank, we developed an integrative PRS combining information from individual scores. Furthermore, in the testing step, we tested the performance of the integrative PRS in another independent set of UK Biobank data of incident cases and controls. RESULTS: Our constructed PRS had improved performance (C statistics: 76.1%) over PRSs constructed by individual benchmark methods (from 69.6% to 74.7%). Furthermore, our new PRS had much higher risk assessment power than family history. The overall net reclassification improvement was 69.0% by adding PRS to the baseline model compared with 12.5% by adding family history. CONCLUSIONS: We developed and validated a new PRS which may improve the utility in predicting the risk of developing PCa. Our innovative method can also be applied to other human diseases to improve risk prediction across multiple outcomes.

Cross-ancestry GWAS meta-analysis identifies six breast cancer loci in African and European ancestry women.

Our study describes breast cancer risk loci using a cross-ancestry GWAS approach. We first identify variants that are associated with breast cancer at P < 0.05 from African ancestry GWAS meta-analysis (9241 cases and 10193 controls), then meta-analyze with European ancestry GWAS data (122977 cases and 105974 controls) from the Breast Cancer Association Consortium. The approach identifies four loci for overall breast cancer risk [1p13.3, 5q31.1, 15q24 (two independent signals), and 15q26.3] and two loci for estrogen receptor-negative disease (1q41 and 7q11.23) at genome-wide significance. Four of the index single nucleotide polymorphisms (SNPs) lie within introns of genes (KCNK2, C5orf56, SCAMP2, and SIN3A) and the other index SNPs are located close to GSTM4, AMPD2, CASTOR2, and RP11-168G16.2. Here we present risk loci with consistent direction of associations in African and European descendants. The study suggests that replication across multiple ancestry populations can help improve the understanding of breast cancer genetics and identify causal variants.

Evaluating Polygenic Risk Scores for Breast Cancer in Women of African Ancestry.

BACKGROUND: Polygenic risk scores (PRSs) have been demonstrated to identify women of European, Asian, and Latino ancestry at elevated risk of developing breast cancer (BC). We evaluated the performance of existing PRSs trained in European ancestry populations among women of African ancestry. METHODS: We assembled genotype data for women of African ancestry, including 9241 case subjects and 10 193 control subjects. We evaluated associations of 179- and 313-variant PRSs with overall and subtype-specific BC risk. PRS discriminatory accuracy was assessed using area under the receiver operating characteristic curve. We also evaluated a recalibrated PRS, replacing the index variant with variants in each region that better captured risk in women of African ancestry and estimated lifetime absolute risk of BC in African Americans by PRS category. RESULTS: For overall BC, the odds ratio per SD of the 313-variant PRS (PRS313) was 1.27 (95% confidence interval [CI] = 1.23 to 1.31), with an area under the receiver operating characteristic curve of 0.571 (95% CI = 0.562 to 0.579). Compared with women with average risk (40th-60th PRS percentile), women in the top decile of PRS313 had a 1.54-fold increased risk (95% CI = 1.38-fold to 1.72-fold). By age 85 years, the absolute risk of overall BC was 19.6% for African American women in the top 1% of PRS313 and 6.7% for those in the lowest 1%. The recalibrated PRS did not improve BC risk prediction. CONCLUSION: The PRSs stratify BC risk in women of African ancestry, with attenuated performance compared with that reported in European, Asian, and Latina populations. Future work is needed to improve BC risk stratification for women of African ancestry.

Racial Distribution and Characterization of Pelvic Organ Prolapse in a Hospital-Based Subspecialty Clinic.

OBJECTIVES: Prior literature has suggested a decreased prevalence of pelvic organ prolapse (POP) in Black women. We sought to describe POP rates by race, investigate whether specific types of prolapse differ based on race, and investigate the role of uterine weight and fibroids on POP. METHODS: We conducted a retrospective cohort study of new patients seen between April 2017 and April 2019 at a tertiary urogynecology clinic. Variables collected included POP quantification, race, age, smoking history, medical history, gravity, parity, vaginal delivery, hysterectomy, fibroids, and uterine weight. χ2 tests were used to compare the proportions of types of POP between Black and non-Black women. Binary and ordinal logistic regression tested the association between types of prolapse and race, adjusting for covariates. RESULTS: Nine hundred thirty-six patients were identified by ICD codes, 768 met inclusion criteria. There were 85.3% of the women identified as non-Black and 14.7% identified as Black. There were 39.8% of the Black women that had a fibroid diagnosis compared with 20.8% of non-Black women (P < 0.001). Black women had a higher median uterine weight, 112.2 g versus 56 g (P = 0.002), and median fibroid size, 3.4 cm versus 1.92 cm (P = 0.0001). 56.9% of women presented with anterior prolapse. No difference was found in POP type between Black and non-Black women after adjusting for age, body mass index, parity, and delivery route (P = 0.45). CONCLUSIONS: Black women had increased body mass index, rates of comorbidities (diabetes and hypertension), higher uterine weight and fibroid size than non-Black women in our study. However, there was no significant difference in POP type based on race.

Daily Aspirin Use Does Not Impact Clinical Outcomes in Patients With Inflammatory Bowel Disease.

BACKGROUND: Although several studies have associated the use of nonsteroidal anti-inflammatory drugs with disease flares in patients with inflammatory bowel disease (IBD), little is known about the impact of daily aspirin use on clinical outcomes in patients with IBD. METHODS: We conducted a retrospective analysis of a prospectively collected registry of patients with IBD from May 2008 to June 2015. Patients with any disease activity with daily aspirin use were matched 1:4 to controls by age, sex, disease, disease location, and presence of cardiac comorbidity. Patients with at least 18 months of follow-up were included in the final analysis. The primary outcomes of interest were having an IBD-related hospitalization, IBD-related surgery, and requiring corticosteroids during the follow-up period. RESULTS: A total of 764 patients with IBD were included in the analysis, of which 174 patients were taking aspirin. There was no statistical difference in age, gender, diagnosis (Crohn's disease vs ulcerative colitis), disease duration, Charlson Comorbidity Index, smoking status, medication usage, or baseline C-reactive protein between groups. After controlling for covariables and length of follow-up in the entire population, aspirin use was not associated with a risk of being hospitalized for an IBD-related complication (odds ratio [OR], 1.46; P = 0.10), corticosteroid use (OR, 0.99; P = 0.70), or having an IBD-related surgery (OR, 0.99; P = 0.96). CONCLUSION: In this single-center analysis, aspirin use did not impact major clinical outcomes in patients with IBD. Although the effect of aspirin use on mucosal inflammation was not directly assessed in this study, these findings support the safety of daily aspirin use in this population.

Penalized Fieller's confidence interval for the ratio of bivariate normal means.

Constructing a confidence interval for the ratio of bivariate normal means is a classical problem in statistics. Several methods have been proposed in the literature. The Fieller method is known as an exact method, but can produce an unbounded confidence interval if the denominator of the ratio is not significantly deviated from 0; while the delta and some numeric methods are all bounded, they are only first-order correct. Motivated by a real-world problem, we propose the penalized Fieller method, which employs the same principle as the Fieller method, but adopts a penalized likelihood approach to estimate the denominator. The proposed method has a simple closed form, and can always produce a bounded confidence interval by selecting a suitable penalty parameter. Moreover, the new method is shown to be second-order correct under the bivariate normality assumption, that is, its coverage probability will converge to the nominal level faster than other bounded methods. Simulation results show that our proposed method generally outperforms the existing methods in terms of controlling the coverage probability and the confidence width and is particularly useful when the denominator does not have adequate power to reject being 0. Finally, we apply the proposed approach to the interval estimation of the median response dose in pharmacology studies to show its practical usefulness.

Discrimination of benign from malignant breast lesions in dense breasts with model-based analysis of regions-of-interest using directional diffusion-weighted images.

BACKGROUND: There is an increasing interest in non-contrast-enhanced magnetic resonance imaging (MRI) for detecting and evaluating breast lesions. We present a methodology utilizing lesion core and periphery region of interest (ROI) features derived from directional diffusion-weighted imaging (DWI) data to evaluate performance in discriminating benign from malignant lesions in dense breasts. METHODS: We accrued 55 dense-breast cases with 69 lesions (31 benign; 38 cancer) at a single institution in a prospective study; cases with ROIs exceeding 7.50 cm2 were excluded, resulting in analysis of 50 cases with 63 lesions (29 benign, 34 cancers). Spin-echo echo-planar imaging DWI was acquired at 1.5 T and 3 T. Data from three diffusion encoding gradient directions were exported and processed independently. Lesion ROIs were hand-drawn on DWI images by two radiologists. A region growing algorithm generated 3D lesion models on augmented apparent-diffusion coefficient (ADC) maps and defined lesion core and lesion periphery sub-ROIs. A lesion-core and a lesion-periphery feature were defined and combined into an overall classifier whose performance was compared to that of mean ADC using receiver operating characteristic (ROC) analysis. Inter-observer variability in ROI definition was measured using Dice Similarity Coefficient (DSC). RESULTS: The region-growing algorithm for 3D lesion model generation improved inter-observer variability over hand drawn ROIs (DSC: 0.66 vs 0.56 (p < 0.001) with substantial agreement (DSC > 0.8) in 46% vs 13% of cases, respectively (p < 0.001)). The overall classifier improved discrimination over mean ADC, (ROC- area under the curve (AUC): 0.85 vs 0.75 and 0.83 vs 0.74 respectively for the two readers). CONCLUSIONS: A classifier generated from directional DWI information using lesion core and lesion periphery information separately can improve lesion discrimination in dense breasts over mean ADC and should be considered for inclusion in computer-aided diagnosis algorithms. Our model-based ROIs could facilitate standardization of breast MRI computer-aided diagnostics (CADx).