Apparent diffusion coefficient measurements in diffusion-weighted magnetic resonance imaging of the anterior mediastinum: inter-observer reproducibility of five different methods of region-of-interest positioning

Development and validation of ADC-based nomogram model for predicting the prognostic factors in preoperative clinical early-stage cervical cancer patients

PURPOSE

To investigate the feasibility of ADC-based nomogram models for predicting cervical cancer (CC) subtype, lymphovascular space invasion (LVSI) and lymph node metastases (LNM) status in preoperative clinical early-stage CC patients.

MATERIALS AND METHODS

A total of 535 CC patients from three independent centers [center A (n = 251) for model training, and centers B (n = 193) and C (n = 91) for external validation] were included. Volumetric ADC histogram metrics (volume, minADC, meanADC, maxADC, skewness, kurtosis, entropy, P10_ADC, P25_ADC, P50_ADC, P75_ADC, and P90_ADC) derived the whole-tumor were calculated. Univariate and multivariate analyses were used to screen the independent predictors and develop nomogram models, with the area under the receiver operating characteristic curve (AUC) for predicting performance estimation.

RESULTS

In differentiating adenosquamous carcinoma (ASC)/adenocarcinoma (AC) from squamous cell carcinoma (SCC), the independent predictors of P25_ADC, SCC antigen (SCC-Ag), and CA199 constructed the nomogram_1 model, with AUCs of 0.900 and 0.873 in training and validation sets, respectively. In differentiating AC from ASC, the independent predictors of P50_ADC and SCC-Ag constructed the nomogram_2 model, with AUCs of 0.837 and 0.829 in training and validation sets, respectively. Tumor volume is the only independent predictor of LVSI(+) and LNM(+), with AUCs of 0.608 and 0.694 in the training set, and 0.553 and 0.656 in the validation set, respectively.

CONCLUSION

The ADC-based nomogram models can effectively predict the CC subtypes, but might be insufficient in predicting the LVSI and LNM status in preoperative clinical early-stage patients.

Development of a nomogram based on whole-tumor multiparametric MRI histogram analysis to predict deep myometrial invasion in stage I endometrioid endometrial carcinoma preoperatively

BACKGROUND

The depth of myometrial invasion determines whether International Federation of Gynecology and Obstetrics stage I endometrioid endometrial carcinoma (EEC) patients undergo lymph node dissection. However, subjective evaluation results relying on magnetic resonance imaging (MRI) are not always satisfactory.

PURPOSE

To develop a nomogram based on whole-volume tumor MRI histogram parameters to preoperatively predict deep myometrial invasion (DMI) in patients with stage I EEC.

MATERIAL AND METHODS

This retrospective analysis included 131 EEC patients and a training/validation cohort of 92/39 patients at a 7:3 ratio. The histogram parameters were obtained from multiple sequences (ADC mapping and T2-weighted imaging) within volumes of interest. Univariate analysis, least absolute shrinkage and selection operator (LASSO) regression, and multivariate logistic regression were used for feature selection. The performance of clinical model, histogram model, and histogram nomogram was evaluated by calculating the area under the receiver operating characteristic curve (AUC).

RESULTS

Age and two morphological features (maximum anteroposterior tumor diameter on sagittal T2-weighted images [APsag] and the tumor area ratio [TAR]) were selected to construct the clinical model. Five histogram parameters were selected for the creation of the histogram model. The nomogram, which combines the histogram parameters, age, APsag, and TAR, achieved the highest AUCs in both the training and validation cohorts (nomogram vs. histogram vs. clinical model: 0.973 vs. 0.871 vs. 0.934 [training] and 0.972 vs. 0.870 vs. 0.928 [validation]).

CONCLUSION

The MR histogram nomogram can help predict the DMI of patients with stage I EEC preoperatively, assisting physicians in the development of personalized treatment strategies.

Whole-tumor apparent diffusion coefficient histogram analysis for preoperative risk stratification in endometrial endometrioid adenocarcinoma

OBJECTIVE

To investigate the application of whole-tumor apparent diffusion coefficient (ADC) histogram metrics for preoperative risk stratification in endometrial endometrioid adenocarcinoma (EEA).

METHODS

Preoperative MRI of 502 EEA patients were retrospectively analyzed. Whole tumor ADC histogram analysis was performed with regions of interest drawn on all tumor slices of diffusion-weighted imaging scans. Risk stratification was based on ESMO-ESTRO-ESP guidelines: low-, intermediate-, high-intermediate-, and high-risk. Univariable analysis was used to compare ADC histogram metrics (tumor volume, minADC, maxADC, and meanADC; 10th, 25th, 50th, 75th, and 90th percentiles of ADC [recorded as P10, P25, P50, P75, and P90 ADC, respectively]; skewness; and kurtosis) between different risk EEAs, and multivariable logistic regression analysis to determine the optimal metric or combined model for risk stratifications. Receiver operating characteristic curve analysis with the area under the curve (AUC) was used for diagnostic performance evaluation.

RESULTS

A decreasing tendency in multiple ADC values was observed from the low- to high-intermediate-risk EEAs. The (low + intermediate)-risk EEAs and low-risk EEAs had significantly smaller tumor volumes and higher minADCs, meanADCs, P10, P25, P50, P75, and P90 ADCs than the (high-intermediate + high)-risk EEAs and non-low-risk EEAs (all P < 0.05), respectively. The combined models of the (meanADC + volume) and the (P75 ADC + volume) yielded the largest AUCs of 0.775 and 0.780 in identifying the (low + intermediate)- and the low-risk EEAs from the other EEAs, respectively.

CONCLUSION

Whole-tumor ADC histogram metrics might be helpful for preoperatively identifying low- and (low + intermediate)-risk EEAs, facilitating personalized therapeutic planning.

Intratumoral Heterogeneity of Fibrosarcoma Xenograft Models: Whole-Tumor Histogram Analysis of DWI and IVIM

RATIONAL AND OBJECTIVE

To explore the correlations of histogram parameters from diffusion-weighted imaging (DWI) and intravoxel incoherent motion (IVIM) with the heterogeneous features in a nude mouse model of fibrosarcoma.

MATERIALS AND METHODS

A total of 44 fibrosarcoma xenograft models were established by inoculating HT-1080 cells on the right thigh of mice and subjected tumors to DWI and IVIM imaging with 3.0 T MRI. Whole-tumor histogram parameters were calculated on apparent diffusion coefficient (ADC), pure diffusion coefficient (D), pseudo-diffusion coefficient (D*), and perfusion fraction (f). Heterogeneous features, including necrosis rate, cell density, Ki-67 labeling index (LI), and microvascular density (MVD) were measured. Intraclass correlation coefficients (ICC), Pearson or Spearman correlation tests, and receiver operating characteristics (ROC) were performed.

RESULTS

The 90th percentile, skewness and kurtosis of ADC and D histograms showed correlations with necrosis rate, and the highest correlation coefficient was found for D90th (r = 0.485). ADC and D histogram parameters showed correlations with cell density and Ki-67 LI; D90th showed the highest correlation coefficient with cell density (r = -0.504); and Dmedian showed the most significant correlation with Ki-67 LI (r = -0.525). D*skewness, D*kurtosis, D*90th, fmean, and fmedian showed correlations with MVD. ADC90th, ADCskewness, ADCkurtosis, D90th, and Dskewness showed significant differences between the low necrosis and high necrosis groups, and the combination model showed the best diagnostic ability (AUC = 0.882), with 97% sensitivity, and 72.7% specificity.

CONCLUSION

Whole-tumor histogram parameters of DWI and IVIM were correlated with heterogeneous features in nude murine models of fibrosarcoma.

Volumetric apparent diffusion coefficient (ADC) histogram metrics as imaging biomarkers for pretreatment predicting response to fertility-sparing treatment in patients with endometrial cancer

OBJECTIVES

To investigate the feasibility of volumetric apparent diffusion coefficient (ADC) histogram analysis for prediction of fertility-sparing treatment (FST) response in patients with endometrial cancer (EC).

METHODS

Pretreatment data of 54 EC patients with FST were retrospectively analyzed. Treatment response at each follow-up was pathologically evaluated. The associations of ADC histogram metrics (volume, minADC, maxADC, meanADC; 10th, 25th, 50th, 75th and 90th ADC percentiles; skewness; kurtosis) and baseline clinical characteristics with complete response (CR) at the second and third follow-ups, two-consecutive CR, and recurrence at the final follow-up were evaluated by uni- and multivariable logistic regression analysis. Receiver operating characteristic (ROC) curve analysis was used for diagnostic performance evaluation.

RESULTS

Compared with non-CR patients, CR patients had significantly higher minADC and 10th and 25th ADC percentiles at the second follow-up (P = 0.008, 0.039, and 0.034, respectively) and higher minADC, older age, lower HE4 level, and higher overweight rate at the third follow-up (P = 0.001, 0.040, 0.021, and 0.004, respectively). Patients with two-consecutive CR had a significantly higher minADC than those without (P = 0.018). There was no association between ADC metrics or clinical characteristics and recurrence (all P > 0.05). MinADC yielded the largest AUC in predicting CR (0.688 and 0.735 at the second and third follow-up, respectively) and the presence of two-consecutive CR (0.753). When combined with patient age and HE4 level, the prediction of CR could be further improved at the third follow-up, with an AUC of 0.786.

CONCLUSION

Pretreatment minADC could be a potential imaging biomarker for predicting FST response. Clinical characteristics may have incremental value to minADC in predicting CR.

Apparent Diffusion Coefficient Value as a Biomarker for Detecting Muscle-Invasive and High-Grade Bladder Cancer: A Systematic Review

Background: Several studies have investigated the potential role of the apparent diffusion coefficient (ADC) value of diffusion-weighted magnetic resonance imaging as a biomarker of high-grade and invasive bladder cancer. Methods: PubMed and the Cochrane Library were systematically searched in September 2021 to extract studies that evaluated the associations between ADC values, pathological T stage, and histological grade bladder cancers. The diagnostic performance of ADC values in detecting muscle-invasive bladder cancer (MIBC) and high-grade disease was systematically reviewed. Results: Six studies were included in this systematic review. MIBC showed significantly lower ADC values than non-muscle-invasive bladder cancer (NMIBC) in all six studies. The median (range) sensitivity, specificity, and area under the curve (AUC) of ADC values to detect MIBC among the four eligible studies were 73.5% (68.8–90.0%), 79.9% (66.7–84.4%), and 0.762 (0.730–0.884), respectively. Similarly, high-grade disease showed significantly lower ADC values than did low-grade disease in all four eligible studies. The median (range) sensitivity, specificity, and AUC of ADC values for detecting high-grade disease among the three eligible studies were 75.0% (73.0–76.5%), 95.8% (76.2–100%), and 0.902 (0.804–0.906), respectively. Conclusions: The ADC value is a non-invasive diagnostic biomarker for discriminating muscle-invasive and high-grade bladder cancer.

Repeatability and reproducibility of MRI apparent diffusion coefficient applied on four different regions of interest for patients with axial spondyloarthritis and healthy volunteers scanned twice within a week

Objectives:

The apparent diffusion coefficient (ADC) may be used as a biomarker for diagnosis and/or monitoring treatment response in patients with axial spondyloarthritis (axSpA), but this requires reliable ADC measurements. This study assessed test–retest repeatability and reproducibility of ADC measurements using four different region of interest (ROI) settings.

Methods:

In this prospective study, the sacroiliac joints (SIJs) of 25 patients with axSpA and 24 age- and sex-matched healthy volunteers were imaged twice at a mean interval of 6.8 days in a 1.5 T scanner using, multishot echoplanar diffusion-weighted sequences. ADCs at four ROI settings were assessed: 5 mm and 10 mm anatomic band-shaped, 15 mm linear, and 40 mm² circular.

Results:

Intraclass correlation coefficient (ICC) assessments showed that the interstudy repeatability was good for median ADC (ADC_med) and 95th-percentile ADC (ADC₉₅) measurements in patients with axSpA (0.77–0.83 and 0.75–0.83, respectively), but poor-to-moderate in healthy subjects (0.27–0.55 and 0.13–0.37, respectively). For all ROI settings, intrareader reproducibility was excellent for ADC_med-measurements (ICC:0.85–0.99) and moderate-to-excellent for ADC₉₅ measurements (ICC:0.68–0.96). The 5 mm ROI had the least estimated bias and highest level of agreement on Bland–Altman plots. The interreader reproducibility was moderate (ICC:0.71). The 15 mm linear ROI produced significantly greater ADC_med and ADC₉₅ measurements than all other ROI settings (p < 0.01–0.02), except for the circular ROI ADC₉₅ measurements.

Conclusion:

ROI settings influence ADC measurements. Interstudy repeatability of SIJ ADC measurements is independent of ROI settings. However, the 5 mm ROI showed the least bias and random error and seems preferable.

Advances in knowledge:

ADC measurements are affected by ROI settings, and this should be taken into account when assessing ADC maps.

Peripheral nerve sheath tumor: differentiation of malignant from benign tumors with conventional and diffusion-weighted MRI

OBJECTIVES

To evaluate potential of conventional MRI and diffusion-weighted imaging (DWI) for differentiating malignant from benign peripheral nerve sheath tumors (PNSTs).

METHODS

Eighty-seven cases of malignant or benign PNSTs in the trunk or extremities that underwent conventional MRI with contrast enhancement, DWI, and pathologic confirmation between Sep. 2014 and Dec. 2017 were identified. Of these, 55 tumors of uncertain nature on MRI were included. Tumor size, signal, and morphology were reviewed on conventional MRI, and apparent diffusion coefficient (ADC) values of solid enhancing portions were measured from DWI. Patient demographics, MRI features, and ADC values were compared between benign and malignant tumors, and robust imaging findings for malignant peripheral nerve sheath tumors (MPNSTs) were identified using multivariable models.

RESULTS

A total of 55 uncertain tumors consisted of 18 malignant and 37 benign PNSTs. On MRI, tumor size, margin, perilesional edema, and presence of split fat, fascicular, and target signs were significantly different between groups (p < 0.05), as were mean and minimum ADC values (p = 0.002, p < 0.0001). Most inter-reader agreement was moderate to excellent (κ value, 0.45-1.0). The mean ADC value and absence of a split fat sign were identified as being associated with MPNSTs (odds ratios = 13.19 and 25.67 for reader 1; 49.05 and 117.91 for reader 2, respectively). The C-indices obtained by combining these two findings were 0.90 and 0.95, respectively.

CONCLUSIONS

Benign and malignant PNSTs showed different features on MRI and DWI. A combination of mean ADC value and absence of split fat was excellent for discriminating malignant from benign PNSTs.

KEY POINTS

• It is important to distinguish between malignant peripheral nerve sheath tumors (MPNSTs) and benign peripheral nerve sheath tumors (BPNSTs) to ensure an appropriate treatment plan. • On conventional MRI and diffusion-weighted imaging (DWI), MPNSTs and BPNSTs showed significant differences in tumor size, margin, presence of perilesional edema, and absence of split fat, fascicular, and target signs. • Absence of a split fat sign and mean apparent diffusion coefficient (ADC) values were robust imaging findings distinguishing MPNSTs from BPNSTs, with a C-index of > 0.9.

Risk stratification of thymic epithelial tumors by using a nomogram combined with radiomic features and TNM staging

OBJECTIVES

To construct and validate a nomogram model that integrated the CT radiomic features and the TNM staging for risk stratification of thymic epithelial tumors (TETs).

METHODS

A total of 136 patients with pathology-confirmed TETs who underwent CT examination were collected from two institutions. According to the WHO pathological classification criteria, patients were classified into low-risk and high-risk groups. The TNM staging was determined in terms of the 8th edition AJCC/UICC staging criteria. LASSO regression was performed to extract the optimal features correlated to risk stratification among the 704 radiomic features calculated. A nomogram model was constructed by combining the Radscore and the TNM staging. The clinical performance was evaluated by ROC analysis, calibration curve, and decision curve analysis (DCA). The Kaplan-Meier (KM) analysis was employed for survival analysis.

RESULTS

Five optimal features identified by LASSO regression were employed to calculate the Radscore correlated to risk stratification. The nomogram model showed a better performance in both training cohort (AUC = 0.84, 95%CI 0.75-0.91) and external validation cohort (AUC = 0.79, 95%CI 0.69-0.88). The calibration curve and DCA analysis indicated a better accuracy of the nomogram model for risk stratification than either Radscore or the TNM staging alone. The KM analysis showed a significant difference between the two groups stratified by the nomogram model (p = 0.02).

CONCLUSIONS

A nomogram model that integrated the radiomic signatures and the TNM staging could serve as a reliable model of risk stratification in predicting the prognosis of patients with TETs.

KEY POINTS

• The radiomic features could be associated with the TET pathophysiology. • TNM staging and Radscore could independently stratify the risk of TETs. • The nomogram model is more objective and more comprehensive than previous methods.

Incremental Value of Magnetic Resonance Imaging in Further Characterizing Hypodense Mediastinal and Paracardiac Lesions Identified on Computed Tomography

Mediastinal and paracardiac lesions are usually first diagnosed on a chest radiograph or echocardiogram. Often, a computed tomography is obtained to further delineate these lesions. CT may be suboptimal for evaluation of enhancement characteristics and direct extension into the adjacent mediastinal structures. With its intrinsic superior soft-tissue characterization, magnetic resonance imaging (MRI) can better delineate these lesions, their internal tissue characteristics, and identify adhesion/invasion into adjacent structures. This pictorial essay provides a brief synopsis of the key MRI sequences and their utility in further characterizing mediastinal and paracardiac lesions.

Nonsuppressing normal thymus on chemical-shift MR imaging and anterior mediastinal lymphoma: differentiation with diffusion-weighted MR imaging by using the apparent diffusion coefficient

OBJECTIVES

To prospectively evaluate usefulness of the apparent diffusion coefficient (ADC) in differentiating anterior mediastinal lymphoma from nonsuppressing normal thymus on chemical-shift MR, and to look at the relationship between patient age and ADC.

METHODS

Seventy-three young subjects (25 men, 48 women; age range, 9-29 years), who underwent chemical-shift MR and diffusion-weighted MR were divided into a normal thymus group (group A, 40 subjects), and a lymphoma group (group B, 33 patients). For group A, all subjects had normal thymus with no suppression on opposed-phase chemical-shift MR. Two readers measured the signal intensity index (SII) and ADC. Differences in SII and ADC between groups were tested using t-test. ADC was correlated with age using Pearson correlation coefficient.

RESULTS

Mean SII±standard deviation was 2.7±1.8% for group A and 2.2±2.4% for group B, with no significant difference between groups (P=.270). Mean ADC was 2.48±0.38x10^-3mm²/s for group A and 1.24±0.23x10^-3mm²/s for group B. A significant difference between groups was found (P<.001), with no overlap in range. Lastly, significant correlation was found between age and ADC (r=0.935, P<.001) in group A.

CONCLUSIONS

ADC of diffusion-weighted MR is a noninvasive and accurate parameter for differentiating lymphoma from nonsuppressing thymus on chemical-shift MR in young subjects.

KEY POINTS

• SII cannot differentiate mediastinal lymphoma from nonsuppressing normal thymus at visual assessment • ADC is useful for distinguishing nonsuppressing normal thymus from mediastinal lymphoma • ADC is more accurate than transverse-diameter and surface-area in this discrimination • ADC of normal thymus is age dependent and increases with increasing age.

Diffusion-weighted quantitative MRI of pleural abnormalities: Intra- and interobserver variability in the apparent diffusion coefficient measurements

PURPOSE

To assess intra- and interobserver variability in the apparent diffusion coefficient (ADC) measurements of pleural abnormalities.

MATERIALS AND METHODS

Diffusion-weighted magnetic resonance imaging was performed in 34 patients to characterize pleural abnormalities, with a 1.5T unit at b values of 0/150/500/800 sec/mm² . In two sessions held 3 months apart, on perfusion-free ADC maps, two independent readers measured the ADC of pleural abnormalities (two readings for each reader in each case) using different methods of region-of-interest (ROI) positioning. In three methods, freehand ROIs were drawn within tumor boundaries to encompass the entire lesion on one or more axial slices (whole tumor volume [WTV], three slices observer-defined [TSOD], single-slice [SS]), while in two methods one or more ROIs were placed on the more restricted areas (multiple small round ROI [MSR], one small round ROI [OSR]). Measurement variability between readings by each reader (intraobserver repeatability) and between readers in first reading (interobserver repeatability) were assessed using intraclass correlation coefficient (ICC) and coefficient of variation (CoV). Analysis of variance (ANOVA) was performed to compare ADC values between the different methods. The measurement time of each case for all methods in first reading was recorded and compared between methods and readers.

RESULTS

All methods demonstrated good (MSR, OSR) and excellent (WTV, TSOD, SS) intra- and interreader agreement, with best and worst repeatability in WTV (lower ICC, 0.977; higher CoV, 3.5%) and OSR (lower ICC, 0.625; higher CoV, 22.8%), respectively. The lower 95% confidence interval of ICC resulted in fair to moderate agreement for OSR (up to 0.379) and in excellent agreement for WTV, TSV, and SS (up to 0.918). ADC values of OSR and MSR were significantly lower compared to other methods (P < 0.001). The OSR and SS required less measurement time (10 and 21/22 sec, respectively) compared to the others (P < 0.0001), while the WTV required the longest measurement time (132/134 sec) (P < 0.0001).

CONCLUSION

ADC measurements of pleural abnormalities are repeatable. The SS method has excellent repeatability, similar to WTV, but requires significantly less measurement time. Thus, its use should be preferred in clinical practice.

LEVEL OF EVIDENCE

4 Technical Efficacy: Stage 2 J. MAGN. RESON. IMAGING 2017;46:769-782.