High-field MR spectroscopy in the multiparametric MRI evaluation of breast lesions

Factors influencing the detection rate of fumarate peak in 1H MR spectroscopy of fumarate hydratase-deficient renal cell carcinoma at 3 T MRI

AIM

To identify factors that may be associated with fumarate detection rate in 1H-magnetic resonance spectroscopy (MRS) in fumarate hydratase-deficient renal cell carcinoma (FH-RCC).

MATERIALS AND MEHODS

Between February 2018 and March 2022, 16 FH-RCC patients with 30 lesions underwent 1H-MRS. Detection results were classified as having a detected fumarate peak (n=12), undetected peak (n=10), or technical failure (n=8). Factors including tumour size, tumour location, treatment history, and metastasis status were collected and analysed. A Bayesian logistic regression model was applied to evaluate the association between these factors and the detection result.

RESULTS

Bayesian analysis demonstrated significant associations between fumarate detection results and the following factors: long-axis diameter (odds ratio [OR] of 1.64; 95% confidence interval [CI] of 1.07-2.53), short-axis diameter (OR of 1.90; 95% CI of 1.19-3.06), voxel size (OR of 2.85; 95% CI of 1.70-4.75), treatment history (OR of 0.35; 95% CI of 0.21-0.58), non-metastatic state (OR of 2.45; 95% CI of 1.48-4.06), and lymph node metastasis (OR of 0.35; 95% CI of 0.21-0.58). Technical failure results were associated with factors such as treatment history (OR of 2.59; 95% CI of 1.37-4.66), non-metastatic state (OR of 0.36; 95% CI of 0.19-0.66), and lymph node metastasis (OR of 2.61; 95% CI of 1.39-4.74).

CONCLUSION

Tumour size, treatment history, and metastasis character were associated with the detection of abnormal fumarate accumulation. This finding will serve as a reference for interpreting 1H-MRS results and for selecting suitable scenarios to evaluate FH-RCC.

Diagnostic Performance of Proton Magnetic Resonance Spectroscopy and 18F-Fluorocholine PET to Differentiate Benign From Malignant Breast Lesions

PURPOSE

The aim of this study was to evaluate the diagnostic performance of the proton magnetic resonance spectroscopy (MRS) and 18F-fluorocholine (FCH) PET for suspicious breast findings on conventional imaging (mammography and breast ultrasound).

METHODS

From September 2012 to December 2015, 37 women with 39 breast lesions on conventional imaging were enrolled and underwent proton MRS and FCH PET. The MRS parameters of choline signal-to-noise ratio (SNR), choline integral (I(cho)), and the PET parameters including SUVmax in the prone (SUV1) and supine (SUV2) positions were analyzed. Receiver operating characteristic curves with the area under the curve, sensitivity, and specificity under the optimal cutoff points for the different parameters were determined.

RESULTS

Twenty-three lesions (59%) were malignant, and 16 (41.0%) were benign. The malignant lesions tended to show significantly higher MRS and PET parameters than benign lesions (choline SNR, P = 0.007; I(cho), P = 0.003; SUV1 and SUV2, P < 0.0001). Fair to moderate correlations were noted between the choline SNR and PET parameters (SUV1, Spearman rank correlation coefficient, ρ = 0.477; SUV2, ρ = 0.483), as well as I(cho) and PET parameters (SUV1, ρ = 0.493; SUV2, ρ = 0.549). The SUV2 showed the highest diagnostic performance (area under the curve, 0.918). Using 2.5 as the optimal cutoff point, the SUV2 yields 89.5% sensitivity and 87.5% specificity for differentiating malignant from benign lesions.

CONCLUSION

The MRS parameters were fairly to moderately correlated with FCH PET parameters, and both could differentiate malignant from benign breast lesions with SUV2 showing best diagnostic performance.

Is Elevated Choline on Magnetic Resonance Spectroscopy a Reliable Marker of Breast Lesion Malignancy?

Background

Contemporary magnetic resonance imaging (MRI) of the breast represents a powerful diagnostic modality for cancer detection, with excellent sensitivity and high specificity. Magnetic resonance spectroscopy (MRS) is being explored as an additional tool for improving specificity in breast cancer detection, using multiparametric MRI. The aim of this study was to examine the possibility of 1H-MRS to discriminate malignant from benign breast lesions, using elevated choline (Cho) peak as an imaging biomarker.

Methods

A total of 60 patients were included in this prospective study: 30 with malignant (average age, 55.2 years; average lesion size, 35 mm) and 30 with benign breast lesions (average age, 44.8 years; average lesion size, 20 mm), who underwent multiparametric MRI with multivoxel 3D ¹H-MRS on a 1.5-T scanner in a 3-year period. Three patients with benign breast lesions were excluded from the study. All lesions were histologically verified. Peaks identified on ¹H-MRS were lipid (0.9, 2.3, 2.8, and 5.2 ppm), choline (3.2 ppm), and water peaks (4.7 ppm). Sensitivity and specificity, as well as positive and negative predictive values, were defined using ROC curves. Cohen's Kappa test of inter-test reliability was performed [testing the agreement between ¹H-MRS and histologic finding, and ¹H-MRS and MR mammography (MRM)].

Results

Choline peak was elevated in 24/30 malignant lesions and in 20/27 benign breast lesions. The sensitivity of ¹H-MRS was 0.8, specificity was 0.741, positive predictive value was 0.774, and negative predictive value was 0.769. Area under ROC was 0.77 (CI 0.640-0.871). Inter-test reliability between ¹H-MRS and histologic finding was 0.543 (moderate agreement) and that between ¹H-MRS and MRM was 0.573 (moderate agreement). False-negative findings were most frequently observed in invasive lobular cancers, while false-positive findings were most frequently observed in adenoid fibroadenomas.

Conclusion

Although elevation of the choline peak has a good sensitivity and specificity in breast cancer detection, both are significantly lower than those of multiparametric MRM. Inclusion of spectra located on tumor margins as well as analysis of lipid peaks could aid both sensitivity and specificity. An important ratio of false-positive and false-negative findings in specific types of breast lesions (lobular cancer and adenoid fibroadenoma) suggests interpreting these lesions with a caveat.

Real-time control of respiratory motion: Beyond radiation therapy

Motion management in radiation oncology is an important aspect of modern treatment planning and delivery. Special attention has been paid to control respiratory motion in recent years. However, other medical procedures related to both diagnosis and treatment are likely to benefit from the explicit control of breathing motion. Quantitative imaging - including increasingly important tools in radiology and nuclear medicine - is among the fields where a rapid development of motion control is most likely, due to the need for quantification accuracy. Emerging treatment modalities like focussed-ultrasound tumor ablation are also likely to benefit from a significant evolution of motion control in the near future. In the present article an overview of available respiratory motion systems along with ongoing research in this area is provided. Furthermore, an attempt is made to envision some of the most expected developments in this field in the near future.

Artificial Intelligence for Breast MRI in 2008-2018: A Systematic Mapping Review

OBJECTIVE

The purpose of this study is to review literature from the past decade on applications of artificial intelligence (AI) to breast MRI.

MATERIALS AND METHODS

In June 2018, a systematic search of the literature was performed to identify articles on the use of AI in breast MRI. For each article identified, the surname of the first author, year of publication, journal of publication, Web of Science Core Collection journal category, country of affiliation of the first author, study design, dataset, study aim(s), AI methods used, and, when available, diagnostic performance were recorded.

RESULTS

Sixty-seven studies, 58 (87%) of which had a retrospective design, were analyzed. When journal categories were considered, 36% of articles were identified as being included in the radiology and imaging journal category. Contrast-enhanced sequences were used for most AI applications (n = 50; 75%) and, on occasion, were combined with other MRI sequences (n = 8; 12%). Four main clinical aims were addressed: breast lesion classification (n = 36; 54%), image processing (n = 14; 21%), prognostic imaging (n = 9; 13%), and response to neoadjuvant therapy (n = 8; 12%). Artificial neural networks, support vector machines, and clustering were the most frequently used algorithms, accounting for 66%. The performance achieved and the most frequently used techniques were then analyzed according to specific clinical aims. Supervised learning algorithms were primarily used for lesion characterization, with the AUC value from ROC analysis ranging from 0.74 to 0.98 (median, 0.87) and with that from prognostic imaging ranging from 0.62 to 0.88 (median, 0.80), whereas unsupervised learning was mainly used for image processing purposes.

CONCLUSION

Interest in the application of advanced AI methods to breast MRI is growing worldwide. Although this growth is encouraging, the current performance of AI applications in breast MRI means that such applications are still far from being incorporated into clinical practice.

A straightforward multiparametric quality control protocol for proton magnetic resonance spectroscopy: Validation and comparison of various 1.5 T and 3 T clinical scanner systems

PURPOSE

The aim of this study was to propose and validate across various clinical scanner systems a straightforward multiparametric quality assurance procedure for proton magnetic resonance spectroscopy (MRS).

METHODS

Eighteen clinical 1.5 T and 3 T scanner systems for MRS, from 16 centres and 3 different manufacturers, were enrolled in the study. A standard spherical water phantom was employed by all centres. The acquisition protocol included 3 sets of single (isotropic) voxel (size 20 mm) PRESS acquisitions with unsuppressed water signal and acquisition voxel position at isocenter as well as off-center, repeated 4/5 times within approximately 2 months. Water peak linewidth (LW) and area under the water peak (AP) were estimated.

RESULTS

LW values [mean (standard deviation)] were 1.4 (1.0) Hz and 0.8 (0.3) Hz for 3 T and 1.5 T scanners, respectively. The mean (standard deviation) (across all scanners) coefficient of variation of LW and AP for different spatial positions of acquisition voxel were 43% (20%) and 11% (11%), respectively. The mean (standard deviation) phantom T₂values were 1145 (50) ms and 1010 (95) ms for 1.5 T and 3 T scanners, respectively. The mean (standard deviation) (across all scanners) coefficients of variation for repeated measurements of LW, AP and T₂ were 25% (20%), 10% (14%) and 5% (2%), respectively.

CONCLUSIONS

We proposed a straightforward multiparametric and not time consuming quality control protocol for MRS, which can be included in routine and periodic quality assurance procedures. The protocol has been validated and proven to be feasible in a multicentre comparison study of a fairly large number of clinical 1.5 T and 3 T scanner systems.

Is there a correlation between 3T multiparametric MRI and molecular subtypes of breast cancer?

OBJECTIVES

To test whether 3 T multiparametric magnetic resonance imaging (mMRI) provides information related to molecular subtypes of breast cancer.

METHODS

Women with mammographic or US findings of breast lesions (BI-RADS 4-5) underwent 3 T mMRI (DCE, DWI and MR spectroscopy). The histological type of breast cancer was assessed. Estrogen-receptor (ER), progesterone-receptor (PgR), Ki-67 status and HER-2 expression, assessed by immunohistochemistry (IHC), defined four molecular subtypes: Luminal-A, Luminal-B, HER2-enriched and triple-negative. Non-parametric tests (Kruskal-Wallis, k-sample equality of medians, and Mann-Whitney), logistic regression or ANOVA, and a multivariate analysis were performed to investigate correlations between the four molecular subtypes and mMRI (lesion volume, margins or distribution, enhancement pattern, ADC, type of kinetic curve, and total choline (tCho) signal-to-noise-ratio (SNR)). A ROC analysis was finally performed to test the diagnostic power of a multivariate logistic regression model.

RESULTS

433 patients (453 lesions) were considered. Volume was smaller in Luminal-B and larger in triple-negative tumours compared to the other subtypes combined. Margins were significantly correlated to Luminal-A and Luminal-B. The type of curve was significantly correlated to Luminal-A. ADC values were higher in Luminal-A. tCho SNR was higher in triple-negative tumours. The ROC analysis showed that the area under the curve (AUC) significantly improved when multiple MRI features were used compared to individual parameters.

CONCLUSIONS

A significant correlation was found between some MRI features and molecular subtypes of breast tumours. A multiparametric approach improved the diagnostic power of MRI. However, further research is needed in order to predict the molecular subtype on the sole basis of mMRI.