Repeatability of derived parameters from histograms following non-Gaussian diffusion modelling of diffusion-weighted imaging in a paediatric oncological cohort

Value of multiple models of diffusion-weighted imaging to predict hepatic lymph node metastases in colorectal liver metastases patients

BACKGROUND

About 10%-31% of colorectal liver metastases (CRLM) patients would concomitantly show hepatic lymph node metastases (LNM), which was considered as sign of poor biological behavior and a relative contraindication for liver resection. Up to now, there's still lack of reliable preoperative methods to assess the status of hepatic lymph nodes in patients with CRLM, except for pathology examination of lymph node after resection.

AIM

To compare the ability of mono-exponential, bi-exponential, and stretched-exponential diffusion-weighted imaging (DWI) models in distinguishing between benign and malignant hepatic lymph nodes in patients with CRLM who received neoadjuvant chemotherapy prior to surgery.

METHODS

In this retrospective study, 97 CRLM patients with pathologically confirmed hepatic lymph node status underwent magnetic resonance imaging, including DWI with ten b values before and after chemotherapy. Various parameters, such as the apparent diffusion coefficient from the mono-exponential model, and the true diffusion coefficient, the pseudo-diffusion coefficient, and the perfusion fraction derived from the intravoxel incoherent motion model, along with distributed diffusion coefficient (DDC) and α from the stretched-exponential model (SEM), were measured. The parameters before and after chemotherapy were compared between positive and negative hepatic lymph node groups. A nomogram was constructed to predict the hepatic lymph node status. The reliability and agreement of the measurements were assessed using the coefficient of variation and intraclass correlation coefficient.

RESULTS

Multivariate analysis revealed that the pre-treatment DDC value and the short diameter of the largest lymph node after treatment were independent predictors of metastatic hepatic lymph nodes. A nomogram combining these two factors demonstrated excellent performance in distinguishing between benign and malignant lymph nodes in CRLM patients, with an area under the curve of 0.873. Furthermore, parameters from SEM showed substantial repeatability.

CONCLUSION

The developed nomogram, incorporating the pre-treatment DDC and the short axis of the largest lymph node, can be used to predict the presence of hepatic LNM in CRLM patients undergoing chemotherapy before surgery. This nomogram was proven to be more valuable, exhibiting superior diagnostic performance compared to quantitative parameters derived from multiple b values of DWI. The nomogram can serve as a preoperative assessment tool for determining the status of hepatic lymph nodes and aiding in the decision-making process for surgical treatment in CRLM patients.

The value of intravoxel incoherent motion diffusion-weighted imaging in predicting perineural invasion for resectable gastric cancer: a prospective study

AIM

To investigate the potential of intravoxel incoherent motion (IVIM) diffusion-weighted imaging to predict perineural invasion (PNI) preoperatively in resectable gastric cancer (GC).

MATERIALS AND METHODS

This study prospectively recruited 85 surgically resected GC patients (58 men, 27 women) aged 60.87 ± 10.17 (39-81) years, who underwent IVIM sequence within 1 week before surgery. According to histopathological PNI diagnoses, patients were divided into PNI positive and negative groups. Conventional apparent diffusion coefficient (ADC) and the IVIM parameters, including true diffusion coefficient (D), pseudodiffusion coefficient (D∗), and pseudodiffusion fraction (f), were compared between the two groups. Morphological MRI features were also analysed. Multivariate logistic regression was used to screen independent predictors of PNI. Receiver-operating characteristic curve analyses were preformed to evaluate the efficacy. Spearman's correlation test was performed to analyse the relationship between MRI parameters and PNI.

RESULTS

Tumour thickness and f in PNI-positive group were higher, whereas the ADC, D were lower than those in PNI-negative group (p<0.05). These four parameters correlated with PNI (p<0.05). The D, f, and tumour thickness were independent predictors of PNI. The area under the curve of ADC, D, f, thickness, and the combined parameter (D + f + thickness) were 0.648, 0.745, 0.698, 0.725, and 0.869, respectively. The combined parameter demonstrated higher efficacy than any other parameters (p<0.05).

CONCLUSION

The ADC, D, and f can effectively distinguish PNI status in GC. The D, f, and thickness were independent predictors of PNI.

Predicting histopathological types and molecular subtype of breast tumors: A comparative study using amide proton transfer-weighted imaging, intravoxel incoherent motion and diffusion kurtosis imaging

PURPOSE

To evaluate the predictive performance of multiparameter and histogram features derived from amide proton transfer-weighted imaging (APTWI), intravoxel incoherent motion (IVIM) and diffusion kurtosis imaging (DKI) for histopathological types of breast tumors.

METHODS

Region of interest (ROI) was delineated by outlining the largest slice of the tumor on the false-color images of the DKI, IVIM and APTWI parameters, and extracted the histogram features. Receiver operating characteristic (ROC) curve was used to evaluate the performance of parameters in predicting benign and malignant breast lesions, molecular prognostic biomarkers, lymph node status, and subtypes of breast lesions. The Spearman correlation coefficient was used to determine the correlations between each parameter and clinical-pathological factors.

RESULTS

All 52 breast lesions were enrolled in this prospective study, including 8 benign lesions and 44 breast cancers. To diagnose malignant and benign breast lesions, the value of APT (min) performed best, with the AUC reaching 0.983. According to the different imaging methods, the APTWI performed best. To predict the positive status of ER, PR, Ki67, the value of Dapp (uniformity), Dapp (uniformity), f (entropy) performed best, with the AUC values reaching 0.743, 0.770, 0.848, respectively. For the identification of Luminal B, HER2-enriched, and TNBC breast cancers, Kapp (max), f (kurtosis), and Dapp (uniformity) performed best, with AUC values reaching 0.679, 0.826, 0.771, respectively.

CONCLUSION

This study found the APTWI, IVIM and DKI parameters could diagnose breast cancer. The histogram features of DKI and IVIM, based on tumor heterogeneity, may help to predict breast cancer subtypes.

Application of intravoxel incoherent motion diffusion-weighted imaging for preoperative knowledge of lymphovascular invasion in gastric cancer: a prospective study

PURPOSE

To investigate the potential of intravoxel incoherent motion diffusion-weighted imaging (IVIM) for preoperative prediction of lymphovascular invasion (LVI) in gastric cancer (GC).

METHODS

This study prospectively enrolled 90 patients (62 males, 28 females, 60.79 ± 9.99 years old) who received radical gastrostomy. Abdominal MRI examinations including IVIM were performed within 1 week before surgery. Patients were divided into LVI-positive and -negative group according to pathological diagnosis after surgery. The apparent diffusion coefficient (ADC) and IVIM parameters, including true diffusion coefficient (D), pseudodiffusion coefficient (D*), and pseudodiffusion fraction (f), were compared between the two groups. The relationship between MRI parameters and LVI was studied by Spearman's correlation analysis. Multivariable logistic regression analysis was used to screen independent predictors of LVI. Receiver-operating characteristic curve analyses were applied to evaluate the efficacy.

RESULTS

The ADC, D in LVI-positive group were lower, whereas tumor thickness and f parameter in LVI-positive group were higher than those in LVI-negative group, and they were statistically correlated with LVI (p < 0.05). D, f and tumor thickness were independent risk factors of LVI. The area under the curve of ADC, D, f, thickness, and the combined parameter (D + f + thickness) were 0.667, 0.754, 0.695, 0.792, and 0.876, respectively. The combined parameter demonstrated higher efficacy than any other parameters (p < 0.05).

CONCLUSION

The ADC, D, and f can effectively distinguish LVI status of GC. The D, f and thickness were independent predictors. The combination of the three predictors further improved the efficacy.

Test-retest repeatability of ADC in prostate using the multi b-Value VERDICT acquisition

PURPOSE

VERDICT (Vascular, Extracellular, Restricted Diffusion for Cytometry in Tumours) MRI is a multi b-value, variable diffusion time DWI sequence that allows generation of ADC maps from different b-value and diffusion time combinations. The aim was to assess precision of prostate ADC measurements from varying b-value combinations using VERDICT and determine which protocol provides the most repeatable ADC.

MATERIALS AND METHODS

Forty-one men (median age: 67.7 years) from a prior prospective VERDICT study (April 2016-October 2017) were analysed retrospectively. Men who were suspected of prostate cancer and scanned twice using VERDICT were included. ADC maps were formed using 5b-value combinations and the within-subject standard deviations (wSD) were calculated per ADC map. Three anatomical locations were analysed per subject: normal TZ (transition zone), normal PZ (peripheral zone), and index lesions. Repeated measures ANOVAs showed which b-value range had the lowest wSD, Spearman correlation and generalized linear model regression analysis determined whether wSD was related to ADC magnitude and ROI size.

RESULTS

The mean lesion ADC for b0b1500 had the lowest wSD in most zones (0.18-0.58x10-4 mm2/s). The wSD was unaffected by ADC magnitude (Lesion: p = 0.064, TZ: p = 0.368, PZ: p = 0.072) and lesion Likert score (p = 0.95). wSD showed a decrease with ROI size pooled over zones (p = 0.019, adjusted regression coefficient = -1.6x10-3, larger ROIs for TZ versus PZ versus lesions). ADC maps formed with a maximum b-value of 500 s/mm2 had the largest wSDs (1.90-10.24x10-4 mm2/s).

CONCLUSION

ADC maps generated from b0b1500 have better repeatability in normal TZ, normal PZ, and index lesions.

Deep-learned estimation of uncertainty in measurements of apparent diffusion coefficient from whole-body diffusion-weighted MRI

PURPOSE

To use deep learning to calculate the uncertainty in apparent diffusion coefficient (σADC) voxel-wise measurements to clinically impact the monitoring of treatment response and improve the quality of ADC maps.

MATERIALS AND METHODS

We use a uniquely designed diffusion-weighted imaging (DWI) acquisition protocol that provides gold-standard measurements of σADC to train a deep learning model on two separate cohorts: 16 patients with prostate cancer and 28 patients with mesothelioma. Our network was trained with a novel cost function, which incorporates a perception metric and a b-value regularisation term, on ADC maps calculated by combinations of 2 or 3 b-values (e.g. 50/600/900, 50/900, 50/600, 600/900 s/mm²). We compare the accuracy of the deep-learning based approach for estimation of σADC with gold-standard measurements.

RESULTS

The model accurately predicted the σADC for every b-value combination in both cohorts. Mean values of σADC within areas of active disease deviated from those measured by the gold-standard by 4.3% (range, 2.87-6.13%) for the prostate and 3.7% (range, 3.06-4.54%) for the mesothelioma cohort. We also showed that the model can easily be adapted for a different DWI protocol and field-of-view with only a few images (as little as a single patient) using transfer learning.

CONCLUSION

Deep learning produces maps of σADC from standard clinical diffusion-weighted images (DWI) when 2 or more b-values are available.

B1 -Corrected T1 Mapping in Lung Cancer: Repeatability, Reproducibility, and Identification of Histological Types

BACKGROUND

T1 mapping can potentially quantitatively assess the intrinsic properties of tumors. B₁ correction can reduce the magnetic field inhomogeneity.

PURPOSE

To assess the repeatability and reproducibility of B₁ -corrected T1 mapping for lung cancer and the ability to identify pathological types.

STUDY TYPE

Prospective reproducibility study.

POPULATION

Sixty lung cancer patients (22 with emphysema) with a total of 60 lesions (adenocarcinoma [n = 23], squamous cell carcinoma [n = 19], and small-cell lung cancer [SCLC] [n = 18]).

FIELD STRENGTH/SEQUENCE

A 3 T/B₁ -corrected 3D variable flip angle T1 mapping and free-breathing diffusion-weighted imaging.

ASSESSMENT

Intraobserver, interobserver, and test-retest reproducibility of minimum, maximum, mean, and SD of lung tumor T1 values were assessed. The correlation between mean T1 and apparent diffusion coefficient (ADC) and differences between different histological types of lung cancer were evaluated.

STATISTICAL TESTS

Intraclass correlation coefficients (ICCs), within-subject coefficients of variation (WCVs), Bland-Altman plots, Pearson's correlation coefficient (r), and analysis of variance (ANOVA). A P value <0.05 was considered to be statistically significant.

RESULTS

No significant differences were found in minimum, maximum, mean, and SD T1 values for repeated measurements (intraobserver and interobserver) and repeated examinations (P = 0.103-0.979). All parameters showed good intraobserver, interobserver and test-retest reproducibility (ICC, 0.780-0.978), except the maximum T1 value (ICC, 0.645-0.922). The mean T1 exhibited the best reproducibility and repeatability, with an average difference <6% for repeated measurements, <8% for repeated scans in lung cancer patients, and<10% for repeated scans in those with emphysema. The mean T1 correlated moderately with ADC (r = -0.580, -0.516, and -0.511 for observers A, B, and C). Both mean T1 and mean ADC were significantly different in SCLC patients compared with those in adenocarcinoma and squamous cell carcinoma patients.

DATA CONCLUSION

The mean T1 from B₁ -corrected T1 mapping is a repeatable parameter with the potential to identify histological types of lung cancer and thus may be a promising imaging biomarker for characterizing lung cancer.

EVIDENCE LEVEL

1 TECHNICAL EFFICACY: Stage 2.

Understanding diffusion-weighted MRI analysis: Repeatability and performance of diffusion models in a benign breast lesion cohort

Diffusion-weighted MRI (DWI) is an important tool for oncology research, with great clinical potential for the classification and monitoring of breast lesions. The utility of parameters derived from DWI, however, is influenced by specific analysis choices. The purpose of this study was to critically evaluate repeatability and curve-fitting performance of common DWI signal representations, for a prospective cohort of patients with benign breast lesions. Twenty informed, consented patients with confirmed benign breast lesions underwent repeated DWI (3 T) using: sagittal single-shot spin-echo echo planar imaging, bipolar encoding, TR/TE: 11,600/86 ms, FOV: 180 x 180 mm, matrix: 90 x 90, slices: 60 x 2.5 mm, iPAT: GRAPPA 2, fat suppression, and 13 b-values: 0-700 s/mm² . A phase-reversed scan (b = 0 s/mm² ) was acquired for distortion correction. Voxel-wise repeat-measures coefficients of variation (CoVs) were derived for monoexponential (apparent diffusion coefficient [ADC]), biexponential (intravoxel incoherent motion: f, D, D*) and stretched exponential (α, DDC) across the parameter histograms for lesion regions of interest (ROIs). Goodness-of-fit for each representation was assessed by Bayesian information criterion. The volume of interest (VOI) definition was repeatable (CoV 13.9%). Within lesions, and across both visits and the cohort, there was no dominant best-fit model, with all representations giving the best fit for a fraction of the voxels. Diffusivity measures from the signal representations (ADC, D, DDC) all showed good repeatability (CoV < 10%), whereas parameters associated with pseudodiffusion (f, D*) performed poorly (CoV > 50%). The stretching exponent α was repeatable (CoV < 12%). This pattern of repeatability was consistent over the central part of the parameter percentiles. Assumptions often made in diffusion studies about analysis choices will influence the detectability of changes, potentially obscuring useful information. No single signal representation prevails within or across lesions, or across repeated visits; parameter robustness is therefore a critical consideration. Our results suggest that stretched exponential representation is more repeatable than biexponential, with pseudodiffusion parameters unlikely to provide clinically useful biomarkers.

Analysis of Renal Diffusion-Weighted Imaging (DWI) Using Apparent Diffusion Coefficient (ADC) and Intravoxel Incoherent Motion (IVIM) Models

Analysis of renal diffusion-weighted imaging (DWI) data to derive markers of tissue properties requires careful consideration of the type, extent, and limitations of the acquired data. Alongside data quality and general suitability for quantitative analysis, choice of diffusion model, fitting algorithm, and processing steps can have consequences for the precision, accuracy, and reliability of derived diffusion parameters. Here we introduce and discuss important steps for diffusion-weighted image processing, and in particular give example analysis protocols and pseudo-code for analysis using the apparent diffusion coefficient (ADC) and intravoxel incoherent motion (IVIM) models. Following an overview of general principles, we provide details of optional steps, and steps for validation of results. Illustrative examples are provided, together with extensive notes discussing wider context of individual steps, and notes on potential pitfalls.This publication is based upon work from the COST Action PARENCHIMA, a community-driven network funded by the European Cooperation in Science and Technology (COST) program of the European Union, which aims to improve the reproducibility and standardization of renal MRI biomarkers. This analysis protocol chapter is complemented by two separate chapters describing the basic concepts and experimental procedure.

Renal Diffusion-Weighted Imaging (DWI) for Apparent Diffusion Coefficient (ADC), Intravoxel Incoherent Motion (IVIM), and Diffusion Tensor Imaging (DTI): Basic Concepts

The specialized function of the kidney is reflected in its unique structure, characterized by juxtaposition of disorganized and ordered elements, including renal glomerula, capillaries, and tubules. The key role of the kidney in blood filtration, and changes in filtration rate and blood flow associated with pathological conditions, make it possible to investigate kidney function using the motion of water molecules in renal tissue. Diffusion-weighted imaging (DWI) is a versatile modality that sensitizes observable signal to water motion, and can inform on the complexity of the tissue microstructure. Several DWI acquisition strategies are available, as are different analysis strategies, and models that attempt to capture not only simple diffusion effects, but also perfusion, compartmentalization, and anisotropy. This chapter introduces the basic concepts of DWI alongside common acquisition schemes and models, and gives an overview of specific DWI applications for animal models of renal disease.This chapter is based upon work from the COST Action PARENCHIMA, a community-driven network funded by the European Cooperation in Science and Technology (COST) program of the European Union, which aims to improve the reproducibility and standardization of renal MRI biomarkers. This introduction chapter is complemented by two separate chapters describing the experimental procedure and data analysis.

3.0 T MRI IVIM-DWI for predicting the efficacy of neoadjuvant chemoradiation for locally advanced rectal cancer

Purpose

The purpose of this study was to determine the diagnostic performance of intravoxel incoherent motion (IVIM) on assessing response to neoadjuvant chemoradiation (nCRT) in patients with Locally Advanced Rectal Cancer (LARC).

Methods

50 patients with rectal cancer who underwent magnetic resonance (MR) imaging before and after nCRT, the values of pre-nCRT and post-nCRT IVIM-DWI parameters apparent diffusion coefficient (ADC), diffusion coefficient (D), false diffusion coefficient (D*), and perfusion fraction (f), together with the percentage changes (∆% parametric value) induced by nCRT were calculated. According to the patient's response to nCRT, the patients were divided into pathological complete response (pCR) and non-pCR groups, Good Response (GR) group and Poor Response (PR) group, and the above values were compared between different groups. Univariate and multiple logistic regression analysis were done to investigate the relation between different parameters and patient nCRT. Draw ROC curve according to sensitivity and specificity, and compare its diagnostic efficacy.

Results

There were no significant differences in the baseline data of 50 patients. After nCRT, the ADC and D values for LARC increased significantly (all p < 0.05). The pCR group (n = 9) had higher preD*, pref, postD*, ∆%ADC and ∆%D values than the non-pCR group (n = 41) (all p < 0.05). The GR group (n = 17) exhibited higher post D, ∆%ADC and ∆%D values than the PR group (n = 33) (all p < 0.05). From the results of Logistic regression analysis found that ∆%ADC and ∆%D were significantly correlated with patients' response to nCRT. Based on ROC analysis, ∆%D had a higher area under the curve value than ∆%ADC (p = 0.009) in discriminating the pCR from non-pCR groups.

Conclusions

IVIM-DWI technology may be helpful in identifying the pCR and GR patients to nCRT for LARC.

The value of intravoxel incoherent motion diffusion-weighted imaging in predicting the pathologic response to neoadjuvant chemotherapy in locally advanced esophageal squamous cell carcinoma

OBJECTIVE

To explore the value of intravoxel incoherent motion diffusion-weighted imaging (IVIM-DWI) for the prediction of pathologic response to neoadjuvant chemotherapy (NAC) in locally advanced esophageal squamous cell carcinoma (ESCC).

MATERIAL AND METHODS

Forty patients with locally advanced ESCC who were treated with NAC followed by radical resection were prospectively enrolled from September 2015 to May 2018. MRI and IVIM were performed within 1 week before and 2-3 weeks after NAC, prior to surgery. Parameters including apparent diffusion coefficient (ADC), true diffusion coefficient (D), pseudodiffusion coefficient (D^*), and pseudodiffusion fraction (f) before and after NAC were measured. Pathologic response was evaluated according to the AJCC tumor regression grade (TRG) system. The changes in IVIM values before and after therapy in different TRG groups were assessed. Receiver operating characteristic (ROC) curves analysis was used to determine the best cutoff value for predicting the pathologic response to NAC.

RESULTS

Twenty-two patients were identified as TRG 2 (responders), and eighteen as TRG 3 (non-responders) in pathologic evaluation. The ADC, D, and f values increased significantly after NAC. The post-NAC D and ΔD values of responders were significantly higher than those of non-responders. The area under the curve (AUC) was 0.722 for post-NAC D and 0.859 for ΔD in predicting pathologic response. The cutoff values of post-NAC D and ΔD were 1.685 × 10^-3 mm²/s and 0.350 × 10^-3 mm²/s, respectively.

CONCLUSION

IVIM-DWI may be used as an effective functional imaging technique to predict pathologic response to NAC in locally advanced ESCC.

KEY POINTS

• The optimal cutoff values of post-NAC D and ΔD for predicting pathologic response to NAC in locally advanced ESCC were 1.685 × 10^-3 mm²/s and 0.350 × 10^-3 mm²/s, respectively. • Pathologic response to NAC in locally advanced ESCC was favorable in patients with post-NAC D and ΔD values that were higher than the optimal cutoff values. • IVIM-DWI can potentially be used to preoperatively predict pathologic response to NAC in esophageal carcinoma. Accurate quantification of the D value derived from IVIM-DWI may eventually translate into an effective and non-invasive marker to predict therapeutic efficacy.

Quantitative Imaging in Pediatric Hepatobiliary Disease

Pediatric hepatobiliary imaging is important for evaluation of not only congenital or structural disease but also metabolic or diffuse parenchymal disease and tumors. A variety of ultrasonography and magnetic resonance imaging (MRI) techniques can be used for these assessments. In ultrasonography, conventional ultrasound imaging as well as vascular imaging, elastography, and contrast-enhanced ultrasonography can be used, while in MRI, fat quantification, T2/T2^* mapping, diffusion-weighted imaging, magnetic resonance elastography, and dynamic contrast-enhanced MRI can be performed. These techniques may be helpful for evaluation of biliary atresia, hepatic fibrosis, nonalcoholic fatty liver disease, sinusoidal obstruction syndrome, and hepatic masses in children. In this review, we discuss each tool in the context of management of hepatobiliary disease in children, and cover various imaging techniques in the context of the relevant physics and their clinical applications for patient care.

Perfusion-driven Intravoxel Incoherent Motion (IVIM) MRI in Oncology: Applications, Challenges, and Future Trends

Recent developments in MR hardware and software have allowed a surge of interest in intravoxel incoherent motion (IVIM) MRI in oncology. Beyond diffusion-weighted imaging (and the standard apparent diffusion coefficient mapping most commonly used clinically), IVIM provides information on tissue microcirculation without the need for contrast agents. In oncology, perfusion-driven IVIM MRI has already shown its potential for the differential diagnosis of malignant and benign tumors, as well as for detecting prognostic biomarkers and treatment monitoring. Current developments in IVIM data processing, and its use as a method of scanning patients who cannot receive contrast agents, are expected to increase further utilization. This paper reviews the current applications, challenges, and future trends of perfusion-driven IVIM in oncology.

Test-retest repeatability of the apparent diffusion coefficient in sacroiliac joint MRI in patients with axial spondyloarthritis and healthy individuals

Background

The apparent diffusion coefficient (ADC) may be used as a biomarker to diagnose axial spondyloarthritis (axSpA) and monitor therapeutic response.

Purpose

To measure the repeatability of the ADC in healthy individuals and in patients with axSpA with and without active sacroiliitis in a test–retest set-up, and to correlate ADC to conventional magnetic resonance imaging (MRI) bone marrow edema (BME) scores and clinical findings.

Material and Methods

A total of 25 patients with axSpA and 24 sex- and age-matched healthy individuals were prospectively examined with MRI twice within 10 days. Short tau inversion recovery (STIR), T1-weighted and diffusion-weighted imaging sequences were performed. Mono-exponential ADC maps were based on four b-values: 0; 50; 500; and 800. Inter-study repeatability and intra-reader reproducibility were investigated in subgroups, as were associations with conventional MRI and clinical findings.

Results

The inter-study repeatability for the median ADC was moderate for all individuals (intraclass correlation coefficient [ICC] 0.66); it was good in patients with axSpA (ICC 0.79) and poor in healthy individuals (ICC 0.27). Significant differences in ADC were found between women and men (P = 0.03), and between patients with versus without BME on STIR (P = 0.01). ADC was associated with an MRI BME score and with age in women.

Conclusion

ADC seems to be a repeatable parameter in patients with axSpA but not in healthy individuals. ADC is correlated with MRI sacroiliac joint BME score and with age in women.

Multicenter Repeatability Study of a Novel Quantitative Diffusion Kurtosis Imaging Phantom

Quantitative kurtosis phantoms are sought by multicenter clinical trials to establish accuracy and precision of quantitative imaging biomarkers on the basis of diffusion kurtosis imaging (DKI) parameters. We designed and evaluated precision, reproducibility, and long-term stability of a novel isotropic (i)DKI phantom fabricated using four families of chemicals based on vesicular and lamellar mesophases of liquid crystal materials. The constructed iDKI phantoms included negative control monoexponential diffusion materials to independently characterize noise and model-induced bias in quantitative kurtosis parameters. Ten test-retest DKI studies were performed on four scanners at three imaging centers over a six-month period. The tested prototype phantoms exhibited physiologically relevant apparent diffusion, D_app, and kurtosis, K_app, parameters ranging between 0.4 and 1.1 (×10^-3 mm²/s) and 0.8 and 1.7 (unitless), respectively. Measured kurtosis phantom K_app exceeded maximum fit model bias (0.1) detected for negative control (zero kurtosis) materials. The material-specific parameter precision [95% CI for D_app: 0.013-0.022(×10^-3 mm²/s) and for K_app: 0.009-0.076] derived from the test-retest analysis was sufficient to characterize thermal and temporal stability of the prototype DKI phantom through correlation analysis of inter-scan variability. The present study confirms a promising chemical design for stable quantitative DKI phantom based on vesicular mesophase of liquid crystal materials. Improvements to phantom preparation and temperature monitoring procedures have potential to enhance precision and reproducibility for future multicenter iDKI phantom studies.

The value of DWI in predicting the response to synchronous radiochemotherapy for advanced cervical carcinoma: comparison among three mathematical models

Background

Diffusion weighted imaging(DWI) mode mainly includes intravoxel incoherent motion (IVIM), stretched exponential model (SEM) and Gaussian diffusion model, but it is still unclear which mode is the most valuable in predicting the response to radiochemotherapy for cervical cancer. This study aims to compare the values of three mathematical models in predicting the response to synchronous radiochemotherapy for cervical cancer.

Methods

Eighty-four patients with cervical cancer were enrolled into this study. They underwent DWI examination by using 12 b-values prior to treatment. The imaging parameters were calculated on the basis of IVIM, SEM and Gaussian diffusion models respectively. The imaging parameters derived from three mathematical modes were compared between responders and non-responders groups. The repeatability of each imaging parameter was assessed.

Results

The ADC, D or DDC value was lower in responders than in non-responders groups (P = 0.03, 0.02, 0.01). The α value was higher in responders group than in non-responders group (P = 0.03). DDC had the largest area under curves (AUC) (=0.948) in predicting the response to treatment. The imaging parameters derived from SEM had better repeatability (CCC for DDC and α were 0.969 and 0.924 respectively) than that derived from other exponential models.

Conclusion

Three exponential modes of DWI are useful for predicting the response to radiochemotherapy for cervical cancer, and SEM may be used as a potential optimal model for predicting treatment effect.

Clinical utility of mono-exponential model diffusion weighted imaging using two b-values compared to the bi- or stretched exponential model for the diagnosis of biliary atresia in infant liver MRI

Purpose

To investigate the clinical utility of mono-exponential model diffusion weighted imaging (DWI) using two b-values compared to the bi- or stretched exponential model to differentiate biliary atresia (BA) from non-BA in pediatric liver magnetic resonance imaging (MRI).

Methods

Patients who underwent liver MRI with DWI for suspected BA from November 2017 to September 2018 were retrospectively included and divided into BA and non-BA groups. Laboratory results including γ-glutamyl transferase (γGT) were compared between the two groups using the Mann-Whitney U test and Fisher’s exact test. The hepatic apparent diffusion coefficient (ADC) 10 using ten b-values and ADC 2 using two b-values were obtained from the mono-exponential model. The slow diffusion coefficient (D), fast diffusion coefficient (D*), and perfusion fraction (f) were obtained from the bi-exponential model. The distributed diffusion coefficient (DDC) and heterogeneity index (α) were measured from the stretched exponential model. Parameters were compared between the two groups using a linear mixed model and diagnostic performance was assessed using the area under the curve (AUC) analysis.

Results

For 12 patients in the BA and five patients in the non-BA group, the ADC 10 (median 0.985 ×10⁻³ mm²/s vs. 1.332 ×10⁻³ mm²/s, p = 0.008), ADC 2 (median 0.987 ×10⁻³ mm²/s vs. 1.335 ×10⁻³ mm²/s, p = 0.017), D* (median 33.2 ×10⁻³ mm²/s vs. 55.3 ×10⁻³ mm²/s, p = 0.021), f (median 13.4%, vs. 22.1%, p = 0.009), and DDC (median 0.889 ×10⁻³ mm²/s vs. 1.323 ×10⁻³ mm²/s, p = 0.009) values were lower and the γGT (median 368.0 IU/L vs. 93.5 IU/L, p = 0.02) and α (median 0.699 vs. 0.556, p = 0.023) values were higher in the BA group. The AUC values for γGT (AUC 0.867 95% confidence interval [CI] 0.616–0.984), ADC 10 (AUC 0.963, 95% CI 0.834–0.998), ADC 2 (AUC 0.925, 95% CI 0.781–0.987), f (AUC 0.850, 95% CI 0.686–0.949), and DDC (AUC 0.925, 95% CI 0.781–0.987) were not significantly different, except for the D* and α values.

Conclusion

Patients with BA had lower ADC 10, ADC 2, D*, f, and DDC values and higher γGT and α values than those in the non-BA group. The diagnostic performance of ADC 2 using only two b-values showed excellent diagnostic performance and was not significantly different from that of γGT, ADC 10, f, and DDC for diagnosing BA.

Non-Gaussian models of diffusion weighted imaging for detection and characterization of prostate cancer: a systematic review and meta-analysis

The importance of Diffusion Weighted Imaging (DWI) in prostate cancer (PCa) diagnosis have been widely handled in literature. In the last decade, due to the mono-exponential model limitations, several studies investigated non-Gaussian DWI models and their utility in PCa diagnosis. Since their results were often inconsistent and conflicting, we performed a systematic review of studies from 2012 examining the most commonly used Non-Gaussian DWI models for PCa detection and characterization. A meta-analysis was conducted to assess the ability of each Non-Gaussian model to detect PCa lesions and distinguish between low and intermediate/high grade lesions. Weighted mean differences and 95% confidence intervals were calculated and the heterogeneity was estimated using the I² statistic. 29 studies were selected for the systematic review, whose results showed inconsistence and an unclear idea about the actual usefulness and the added value of the Non-Gaussian model parameters. 12 studies were considered in the meta-analyses, which showed statistical significance for several non-Gaussian parameters for PCa detection, and to a lesser extent for PCa characterization. Our findings showed that Non-Gaussian model parameters may potentially play a role in the detection and characterization of PCa but further studies are required to identify a standardized DWI acquisition protocol for PCa diagnosis.

q-Space Imaging Yields a Higher Effect Gradient to Assess Cellularity than Conventional Diffusion-weighted Imaging Methods at 3.0 T: A Pilot Study with Freshly Excised Whole-Breast Tumors

Purpose

To determine whether q-space imaging (QSI), an advanced diffusion-weighted MRI method, provides a higher effect gradient to assess tumor cellularity than existing diffusion imaging methods, and fidelity to cellularity obtained from histologic analysis.

Materials and Methods

In this prospective study, diffusion-weighted images were acquired from 20 whole-breast tumors freshly excised from participants (age range, 35-78 years) by using a clinical 3.0-T MRI unit. Median and skewness values were extracted from the histogram distributions obtained from QSI, monoexponential model, diffusion kurtosis imaging (DKI), and stretched exponential model (SEM). The skewness from QSI and other diffusion models was compared by using paired t tests and relative effect gradient obtained from correlating skewness values.

Results

The skewness obtained from QSI (mean, 1.34 ± 0.77 [standard deviation]) was significantly higher than the skewness from monoexponential fitting approach (mean, 1.09 ± 0.67; P = .015), SEM (mean, 1.07 ± 0.70; P = .014), and DKI (mean, 0.97 ± 0.63; P = .004). QSI yielded a higher effect gradient in skewness (percentage increase) compared with monoexponential fitting approach (0.26 of 0.74; 35.1%), SEM (0.26 of 0.74; 35.1%), and DKI (0.37 of 0.63; 58.7%). The skewness and median from QSI were significantly correlated with the skewness (ρ = -0.468; P = .038) and median (ρ = -0.513; P = .021) of cellularity from histologic analysis.

Conclusion

QSI yields a higher effect gradient in assessing breast tumor cellularity than existing diffusion methods, and fidelity to underlying histologic structure.Keywords: Breast, MR-Diffusion Weighted Imaging, MR-Imaging, Pathology, Tissue Characterization, Tumor ResponseOnline supplemental material is available for this article.Published under a CC BY 4.0 license.

Histogram analysis of diffusion kurtosis imaging based on whole-volume images of breast lesions

BACKGROUND

Breast diffusion kurtosis imaging (DKI) is a novel MRI technique to assess breast cancer but the effectivity still remains to be improved.

PURPOSE

To investigate the performance of whole-volume histogram parameters derived from a DKI model for differentiating benign and malignant breast lesions.

STUDY TYPE

Retrospective.

POPULATION

In all, 120 patients with breast lesions (62 malignant, 58 benign).

SEQUENCE

DKI sequence with seven b-values (0, 500, 1000, 1500, 2000, 2500, and 3000 s/mm² ) and DWI sequence with two b-values (0 and 1000 s/mm² ) on 3.0T MRI.

ASSESSMENT

Histogram parameters of the DKI model (K and D) and the DWI model (ADC), including the minimum, maximum, mean, percentile values (25th, 50th, 75th, and 95th), standard deviation, kurtosis and skewness, were calculated by two radiologists for the whole lesion volume.

STATISTICAL TESTS

Student's t-test was used to compare malignant and benign lesions. The diagnostic performances were evaluated by receiver operating characteristic (ROC) analysis.

RESULTS

K_max , D_min , and ADC_min had the highest area under the curve (AUC) (0.875, 0.830, and 0.847, respectively), sensitivity (85.5%, 74.2%, and 77.4%, respectively), and accuracy (85.0%, 79.2%, and 81.7%, respectively) in their individual histogram parameter groups, and K_max was found to outperform D_min and ADC_min . ADC histogram parameters (from ADC_min to ADC_sd ) were significantly lower than D histogram parameters in all groups.

DATA CONCLUSION

K_max , D_min , and ADC_min were found to be better metrics than the corresponding average values for differentiating benign from malignant tumors. Histogram parameters derived from the DKI model provided more information and had better diagnostic performance than ADC parameters derived from the DWI model.

LEVEL OF EVIDENCE

3 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2020;51:627-634.

Utility of Multi-Parametric Quantitative Magnetic Resonance Imaging for Characterization and Radiotherapy Response Assessment in Soft-Tissue Sarcomas and Correlation With Histopathology

Purpose: To evaluate repeatability of quantitative multi-parametric MRI in retroperitoneal sarcomas, assess parameter changes with radiotherapy, and correlate pre-operative values with histopathological findings in the surgical specimens. Materials and Methods: Thirty patients with retroperitoneal sarcoma were imaged at baseline, of whom 27 also underwent a second baseline examination for repeatability assessment. 14/30 patients were treated with pre-operative radiotherapy and were imaged again after completing radiotherapy (50.4 Gy in 28 daily fractions, over 5.5 weeks). The following parameter estimates were assessed in the whole tumor volume at baseline and following radiotherapy: apparent diffusion coefficient (ADC), parameters of the intra-voxel incoherent motion model of diffusion-weighted MRI (D, f, D*), transverse relaxation rate, fat fraction, and enhancing fraction after gadolinium-based contrast injection. Correlation was evaluated between pre-operative quantitative parameters and histopathological assessments of cellularity and fat fraction in post-surgical specimens (ClinicalTrials.gov, registration number NCT01902667). Results: Upper and lower 95% limits of agreement were 7.1 and -6.6%, respectively for median ADC at baseline. Median ADC increased significantly post-radiotherapy. Pre-operative ADC and D were negatively correlated with cellularity (r = -0.42, p = 0.01, 95% confidence interval (CI) -0.22 to -0.59 for ADC; r = -0.45, p = 0.005, 95% CI -0.25 to -0.62 for D), and fat fraction from Dixon MRI showed strong correlation with histopathological assessment of fat fraction (r = 0.79, p = 10-7, 95% CI 0.69-0.86). Conclusion: Fat fraction on MRI corresponded to fat content on histology and therefore contributes to lesion characterization. Measurement repeatability was excellent for ADC; this parameter increased significantly post-radiotherapy even in disease categorized as stable by size criteria, and corresponded to cellularity on histology. ADC can be utilized for characterizing and assessing response in heterogeneous retroperitoneal sarcomas.

Can IVIM help predict HCC recurrence after hepatectomy?

PURPOSE

To determine the diagnostic performance of intravoxel incoherent motion (IVIM) parameters to predict tumor recurrence after hepatectomy in patients with hepatitis B virus (HBV)-related hepatocellular carcinoma (HCC).

MATERIALS AND METHODS

One hundred and fifty-seven patients (mean age 52.54 ± 11.32 years, 87% male) with surgically and pathologically confirmed HCC were included. Regions of interests were drawn including the tumors by two independent radiologists. ADC and IVIM-derived parameters (true diffusion coefficient [D]; pseudodiffusion coefficient [D*]; pseudodiffusion fraction [f]) were obtained preoperatively. The Cox proportional hazards model was used to analyze the predictors associated with tumor recurrence after hepatectomy.

RESULTS

Forty-seven of 157 (29.9%) patients experienced tumor recurrence. The multivariate Cox proportional hazards model revealed that a D value < 0.985 × 10^-3 mm²/s (hazard ratio (HR), 0.190; p = 0.023) was a risk factor for tumor recurrence. Additional risk factors included younger age (HR, 0.328; p = 0.034) and higher serum alpha-fetoprotein (AFP) level (HR, 2.079; p = 0.013). Further, receiver operating characteristic (ROC) analysis showed that the area under the curve (AUC) of the obtained Cox regression model improved from 0.68 for the combination of AFP and age alone to 0.724 for the combination of D value, AFP, and age.

CONCLUSION

The D value derived from the IVIM model is a potential biomarker for the preoperative prediction of recurrence after hepatectomy in patients with HCC. When combined with age and AFP levels, D can improve the predictive performance for tumor recurrence.

KEY POINTS

• The recurrence rate of HCC after hepatectomy was higher in patients with ADC, D, and f values that were lower than the optimal cutoff values. • The optimal cutoff values of ADC, D, D*, and f for predicting recurrence in HBV associated HCC were 0.858 × 10^-3 mm²/s, 0.985 × 10^-3 mm²/s, 12.5 × 10^-3 mm²/s, and 23.4%, respectively. • The D value derived from IVIM diffusion-weighted imaging may be a useful biomarker for preoperative prediction of recurrence after hepatectomy in patients with HCC. When combined with age and AFP levels, D can improve the predictive performance for tumor recurrence.

Post-radiotherapy apparent diffusion coefficient (ADC) in children and young adults with high-grade gliomas and diffuse intrinsic pontine gliomas

Objectives: Diffusion-weighted magnetic resonance imaging (DW-MRI) offers potential to monitor response and predict survival in high-grade gliomas (HGG) and diffuse intrinsic pontine gliomas (DIPG). We hypothesized that post-radiotherapy DW-MRI may provide prognostic imaging biomarkers in children and young adults with these tumors. Methods: Patients aged ≤21 years diagnosed between 2005 and 2012 were eligible. The tumor median apparent diffusion coefficient (ADC) and its 5th percentile (C5-ADC) were determined at the first post-radiotherapy scan and at the time of radiological progression. DW-MRI parameters were correlated with survival endpoints, temozolomide use and pseudoprogression, when it occurred. Results: Out of 40 patients (20 HGG, 20 DIPG), 23 had evaluable DW-MRI post-radiotherapy and 25 at radiological progression. There were 6 episodes of pseudoprogression. Hazard ratios (95%CI) for progression-free survival were 0.998 (0.993-1.003) for median ADC and 1.003 (0.996-1.010) for C5-ADC. Hazard ratios (95%CI) for overall survival were 1.0009 (0.996-1.006) for median ADC and 0.998 (0.992-1.004) for C5-ADC. Post-radiotherapy median and C5-ADC values were not significantly different between patients treated with radiotherapy alone versus radiotherapy/temozolomide. The median and C5-ADC values were not significantly different at the time of pseudoprogression compared to those at tumor progression. Conclusions: Post-radiotherapy median ADC and C5-ADC were not prognostic, nor able to differentiate radiosensitization with temozolomide or occurrence of pseudoprogression in this cohort of HGG and DIPG patients. Further exploration of alternative DW parameters, study timepoints or data modeling may contribute to the development of prognostic/predictive imaging biomarkers for children and young adults with HGG or DIPG.

Quantitative Whole-Body Diffusion-Weighted MR Imaging

Whole-body diffusion-weighted MRI has emerged as a powerful diagnostic tool for disease detection and staging mainly used in systemic bone disease. The large field-of-view functional imaging technique highlights cellular tumor and suppresses normal tissue signal, allowing quantification of an estimate of total disease burden, summarized as the total diffusion volume (tDV), as well as global apparent diffusion coefficient (gADC) measurements. Both tDV and gADC have been shown to be repeatable quantitative parameters that indicate tumor heterogeneity and treatment effects, thus potential, noninvasive, imaging biomarkers informing on disease prognosis and therapy response.

Diffusion MRI of cancer: From low to high b-values

Following its success in early detection of cerebral ischemia, diffusion-weighted imaging (DWI) has been increasingly used in cancer diagnosis and treatment evaluation. These applications are propelled by the rapid development of novel diffusion models to extract biologically valuable information from diffusion-weighted MR signals, and significant advances in MR hardware that has enabled image acquisition with high b-values. This article reviews recent technical developments and clinical applications in cancer imaging using DWI, with a special emphasis on high b-value diffusion models. The article is organized in four sections. First, we provide an overview of diffusion models that are relevant to cancer imaging. The model parameters are discussed in relation to three tissue properties-cellularity, vascularity, and microstructures. An emphasis is placed on characterization of microstructural heterogeneity, given its novelty and close relevance to cancer. Second, we illustrate diffusion MR clinical applications in each of the following three categories: 1) cancer detection and diagnosis; 2) cancer grading, staging, and classification; and 3) cancer treatment response prediction and evaluation. Third, we discuss several practical issues, including selection of image acquisition parameters, reproducibility and reliability, motion management, image distortion, etc., that are commonly encountered when applying DWI to cancer in clinical settings. Lastly, we highlight a few ongoing challenges and provide some possible future directions, particularly in the area of establishing standards via well-organized multicenter clinical trials to accelerate clinical translation of advanced DWI techniques to improving cancer care on a large scale. Level of Evidence: 5 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2019;49:23-40.

Characterisation of fibrosis in chemically-induced rat mammary carcinomas using multi-modal endogenous contrast MRI on a 1.5T clinical platform

OBJECTIVES

To determine the ability of multi-parametric, endogenous contrast MRI to detect and quantify fibrosis in a chemically-induced rat model of mammary carcinoma.

METHODS

Female Sprague-Dawley rats (n=18) were administered with N-methyl-N-nitrosourea; resulting mammary carcinomas underwent nine-b-value diffusion-weighted (DWI), ultrashort-echo (UTE) and magnetisation transfer (MT) magnetic resonance imaging (MRI) on a clinical 1.5T platform, and associated quantitative MR parameters were calculated. Excised tumours were histologically assessed for degree of necrosis, collagen, hypoxia and microvessel density. Significance level adjusted for multiple comparisons was p=0.0125.

RESULTS

Significant correlations were found between MT parameters and degree of picrosirius red staining (r > 0.85, p < 0.0002 for ka and δ, r < -0.75, p < 0.001 for T1 and T1s, Pearson), indicating that MT is sensitive to collagen content in mammary carcinoma. Picrosirius red also correlated with the DWI parameter fD* (r=0.801, p=0.0004) and conventional gradient-echo T2* (r=-0.660, p=0.0055). Percentage necrosis correlated moderately with ultrashort/conventional-echo signal ratio (r=0.620, p=0.0105). Pimonidazole adduct (hypoxia) and CD31 (microvessel density) staining did not correlate with any MR parameter assessed.

CONCLUSIONS

Magnetisation transfer MRI successfully detects collagen content in mammary carcinoma, supporting inclusion of MT imaging to identify fibrosis, a prognostic marker, in clinical breast MRI examinations.

KEY POINTS

• Magnetisation transfer imaging is sensitive to collagen content in mammary carcinoma. • Magnetisation transfer imaging to detect fibrosis in mammary carcinoma fibrosis is feasible. • IVIM diffusion does not correlate with microvessel density in preclinical mammary carcinoma.

Liver fibrosis: stretched exponential model outperforms mono-exponential and bi-exponential models of diffusion-weighted MRI

OBJECTIVES

To compare the ability of diffusion-weighted imaging (DWI) parameters acquired from three different models for the diagnosis of hepatic fibrosis (HF).

METHODS

Ninety-five patients underwent DWI using nine b values at 3 T magnetic resonance. The hepatic apparent diffusion coefficient (ADC) from a mono-exponential model, the true diffusion coefficient (D _t ), pseudo-diffusion coefficient (D _p ) and perfusion fraction (f) from a biexponential model, and the distributed diffusion coefficient (DDC) and intravoxel heterogeneity index (α) from a stretched exponential model were compared with the pathological HF stage. For the stretched exponential model, parameters were also obtained using a dataset of six b values (DDC^#, α^#). The diagnostic performances of the parameters for HF staging were evaluated with Obuchowski measures and receiver operating characteristics (ROC) analysis. The measurement variability of DWI parameters was evaluated using the coefficient of variation (CoV).

RESULTS

Diagnostic accuracy for HF staging was highest for DDC^# (Obuchowski measures, 0.770 ± 0.03), and it was significantly higher than that of ADC (0.597 ± 0.05, p < 0.001), D _t (0.575 ± 0.05, p < 0.001) and f (0.669 ± 0.04, p = 0.035). The parameters from stretched exponential DWI and D _p showed higher areas under the ROC curve (AUCs) for determining significant fibrosis (≥F2) and cirrhosis (F = 4) than other parameters. However, D _p showed significantly higher measurement variability (CoV, 74.6%) than DDC^# (16.1%, p < 0.001) and α^# (15.1%, p < 0.001).

CONCLUSIONS

Stretched exponential DWI is a promising method for HF staging with good diagnostic performance and fewer b-value acquisitions, allowing shorter acquisition time.

KEY POINTS

• Stretched exponential DWI provides a precise and accurate model for HF staging. • Stretched exponential DWI parameters are more reliable than D _p from bi-exponential DWI model • Acquisition of six b values is sufficient to obtain accurate DDC and α.

Support vector machine for breast cancer classification using diffusion-weighted MRI histogram features: Preliminary study

BACKGROUND

Diffusion-weighted MRI (DWI) is currently one of the fastest developing MRI-based techniques in oncology. Histogram properties from model fitting of DWI are useful features for differentiation of lesions, and classification can potentially be improved by machine learning.

PURPOSE

To evaluate classification of malignant and benign tumors and breast cancer subtypes using support vector machine (SVM).

STUDY TYPE

Prospective.

SUBJECTS

Fifty-one patients with benign (n = 23) and malignant (n = 28) breast tumors (26 ER+, whereof six were HER2+).

FIELD STRENGTH/SEQUENCE

Patients were imaged with DW-MRI (3T) using twice refocused spin-echo echo-planar imaging with echo time / repetition time (TR/TE) = 9000/86 msec, 90 × 90 matrix size, 2 × 2 mm in-plane resolution, 2.5 mm slice thickness, and 13 b-values.

ASSESSMENT

Apparent diffusion coefficient (ADC), relative enhanced diffusivity (RED), and the intravoxel incoherent motion (IVIM) parameters diffusivity (D), pseudo-diffusivity (D*), and perfusion fraction (f) were calculated. The histogram properties (median, mean, standard deviation, skewness, kurtosis) were used as features in SVM (10-fold cross-validation) for differentiation of lesions and subtyping.

STATISTICAL TESTS

Accuracies of the SVM classifications were calculated to find the combination of features with highest prediction accuracy. Mann-Whitney tests were performed for univariate comparisons.

RESULTS

For benign versus malignant tumors, univariate analysis found 11 histogram properties to be significant differentiators. Using SVM, the highest accuracy (0.96) was achieved from a single feature (mean of RED), or from three feature combinations of IVIM or ADC. Combining features from all models gave perfect classification. No single feature predicted HER2 status of ER + tumors (univariate or SVM), although high accuracy (0.90) was achieved with SVM combining several features. Importantly, these features had to include higher-order statistics (kurtosis and skewness), indicating the importance to account for heterogeneity.

DATA CONCLUSION

Our findings suggest that SVM, using features from a combination of diffusion models, improves prediction accuracy for differentiation of benign versus malignant breast tumors, and may further assist in subtyping of breast cancer.

LEVEL OF EVIDENCE

3 Technical Efficacy: Stage 3 J. Magn. Reson. Imaging 2018;47:1205-1216.

Intravoxel incoherent motion (IVIM) histogram biomarkers for prediction of neoadjuvant treatment response in breast cancer patients

OBJECTIVE

To examine the prognostic capabilities of intravoxel incoherent motion (IVIM) metrics and their ability to predict response to neoadjuvant treatment (NAT). Additionally, to observe changes in IVIM metrics between pre- and post-treatment MRI.

METHODS

This IRB-approved, HIPAA-compliant retrospective study observed 31 breast cancer patients (32 lesions). Patients underwent standard bilateral breast MRI along with diffusion-weighted imaging before and after NAT. Six patients underwent an additional IVIM-MRI scan 12-14 weeks after initial scan and 2 cycles of treatment. In addition to apparent diffusion coefficients (ADC) from monoexponential decay, IVIM mean values (tissue diffusivity D_t, perfusion fraction f_p, and pseudodiffusivity D_p) and histogram metrics were derived using a biexponential model. An additional filter identified voxels of highly vascular tumor tissue (VTT), excluding necrotic or normal tissue. Clinical data include histology of biopsy and clinical response to treatment through RECIST assessment. Comparisons of treatment response were made using Wilcoxon rank-sum tests.

RESULTS

Average, kurtosis, and skewness of pseudodiffusion D_p significantly differentiated RECIST responders from nonresponders. ADC and D_t values generally increased (∼70%) and VTT% values generally decreased (∼20%) post-treatment.

CONCLUSION

D_p metrics showed prognostic capabilities; slow and heterogeneous pseudodiffusion offer poor prognosis. Baseline ADC/D_t parameters were not significant predictors of response. This work suggests that IVIM mean values and heterogeneity metrics may have prognostic value in the setting of breast cancer NAT.

Extracranial Soft-Tissue Tumors: Repeatability of Apparent Diffusion Coefficient Estimates from Diffusion-weighted MR Imaging

Purpose To assess the repeatability of apparent diffusion coefficient (ADC) estimates in extracranial soft-tissue diffusion-weighted magnetic resonance imaging across a wide range of imaging protocols and patient populations. Materials and Methods Nine prospective patient studies and one prospective volunteer study, performed between 2006 and 2016 with research ethics committee approval and written informed consent from each subject, were included in this single-institution study. A total of 141 tumors and healthy organs were imaged twice (interval between repeated examinations, 45 minutes to 10 days, depending the on study) to assess the repeatability of median and mean ADC estimates. The Levene test was used to determine whether ADC repeatability differed between studies. The Pearson linear correlation coefficient was used to assess correlation between coefficient of variation (CoV) and the year the study started, study size, and volumes of tumors and healthy organs. The repeatability of ADC estimates from small, medium, and large tumors and healthy organs was assessed irrespective of study, and the Levene test was used to determine whether ADC repeatability differed between these groups. Results CoV aggregated across all studies was 4.1% (range for each study, 1.7%-6.5%). No correlation was observed between CoV and the year the study started or study size. CoV was weakly correlated with volume (r = -0.5, P = .1). Repeatability was significantly different between small, medium, and large tumors (P < .05), with the lowest CoV (2.6%) for large tumors. There was a significant difference in repeatability between studies-a difference that did not persist after the study with the largest tumors was excluded. Conclusion ADC is a robust imaging metric with excellent repeatability in extracranial soft tissues across a wide range of tumor sites, sizes, patient populations, and imaging protocol variations. Online supplemental material is available for this article.

Extended T2-IVIM model for correction of TE dependence of pseudo-diffusion volume fraction in clinical diffusion-weighted magnetic resonance imaging

The bi-exponential intravoxel-incoherent-motion (IVIM) model for diffusion-weighted MRI (DWI) fails to account for differential T 2 s in the model compartments, resulting in overestimation of pseudodiffusion fraction f. An extended model, T2-IVIM, allows removal of the confounding echo-time (TE) dependence of f, and provides direct compartment T 2 estimates. Two consented healthy volunteer cohorts (n  =  5, 6) underwent DWI comprising multiple TE/b-value combinations (Protocol 1: TE  =  62-102 ms, b  =  0-250 mm-2s, 30 combinations. Protocol 2: 8 b-values 0-800 mm-2s at TE  =  62 ms, with 3 additional b-values 0-50 mm-2s at TE  =  80, 100 ms; scanned twice). Data from liver ROIs were fitted with IVIM at individual TEs, and with the T2-IVIM model using all data. Repeat-measures coefficients of variation were assessed for Protocol 2. Conventional IVIM modelling at individual TEs (Protocol 1) demonstrated apparent f increasing with longer TE: 22.4  ±  7% (TE  =  62 ms) to 30.7  ±  11% (TE  =  102 ms); T2-IVIM model fitting accounted for all data variation. Fitting of Protocol 2 data using T2-IVIM yielded reduced f estimates (IVIM: 27.9  ±  6%, T2-IVIM: 18.3  ±  7%), as well as T 2  =  42.1  ±  7 ms, 77.6  ±  30 ms for true and pseudodiffusion compartments, respectively. A reduced Protocol 2 dataset yielded comparable results in a clinical time frame (11 min). The confounding dependence of IVIM f on TE can be accounted for using additional b/TE images and the extended T2-IVIM model.

Non-invasive quantification of tumour heterogeneity in water diffusivity to differentiate malignant from benign tissues of urinary bladder: a phase I study

OBJECTIVES

To quantify the heterogeneity of the tumour apparent diffusion coefficient (ADC) using voxel-based analysis to differentiate malignancy from benign wall thickening of the urinary bladder.

METHODS

Nineteen patients with histopathological findings of their cystectomy specimen were included. A data set of voxel-based ADC values was acquired for each patient's lesion. Histogram analysis was performed on each data set to calculate uniformity (U) and entropy (E). The k-means clustering of the voxel-wised ADC data set was implemented to measure mean intra-cluster distance (MICD) and largest inter-cluster distance (LICD). Subsequently, U, E, MICD, and LICD for malignant tumours were compared with those for benign lesions using a two-sample t-test.

RESULTS

Eleven patients had pathological confirmation of malignancy and eight with benign wall thickening. Histogram analysis showed that malignant tumours had a significantly higher degree of ADC heterogeneity with lower U (P = 0.016) and higher E (P = 0.005) than benign lesions. In agreement with these findings, k-means clustering of voxel-wise ADC indicated that bladder malignancy presented with significantly higher MICD (P < 0.001) and higher LICD (P = 0.002) than benign wall thickening.

CONCLUSIONS

The quantitative assessment of tumour diffusion heterogeneity using voxel-based ADC analysis has the potential to become a non-invasive tool to distinguish malignant from benign tissues of urinary bladder cancer.

KEY POINTS

• Heterogeneity is an intrinsic characteristic of tumoral tissue. • Non-invasive quantification of tumour heterogeneity can provide adjunctive information to improve cancer diagnosis accuracy. • Histogram analysis and k-means clustering can quantify tumour diffusion heterogeneity. • The quantification helps differentiate malignant from benign urinary bladder tissue.