Enhancing the Kaiser score for lesion characterization in unenhanced breast MRI.

PURPOSE: To adapt the methodology of the Kaiser score, a clinical decision rule for lesion characterization in breast MRI, for unenhanced protocols. METHOD: In this retrospective IRB-approved cross-sectional study, we included 93 consecutive patients who underwent breast MRI between 2021 and 2023 for further work-up of BI-RADS 0, 3-5 in conventional imaging or for staging purposes (BI-RADS 6). All patients underwent biopsy for histologic verification or were followed for a minimum of 12 months. MRI scans were conducted using 1.5 T or 3 T scanners using dedicated breast coils and a protocol in line with international recommendations including DWI and ADC. Lesion characterization relied solely on T2w and DWI/ADC-derived features (such as lesion type, margins, shape, internal signal, surrounding tissue findings, ADC value). Statistical analysis was done using decision tree analysis aiming to distinguish benign (histology/follow-up) from malignant outcomes. RESULTS: We analyzed a total of 161 lesions (81 of them non-mass) with a malignancy rate of 40%. Lesion margins (spiculated, irregular, or circumscribed) were identified as the most important criterion within the decision tree, followed by the ADC value as second most important criterion. The resulting score demonstrated a strong diagnostic performance with an AUC of 0.840, providing both rule-in and rule-out criteria. In an independent test set of 65 lesions the diagnostic performance was verified by two readers (AUC 0.77 and 0.87, kappa: 0.62). CONCLUSIONS: We developed a clinical decision rule for unenhanced breast MRI including lesion margins and ADC value as the most important criteria, achieving high diagnostic accuracy.

Cancer detection rate of breast-MR in supplemental screening after negative mammography in women with dense breasts. Preliminary results of the MA-DETECT-Study after 200 participants.

BACKGROUND: Due to increased cancer detection rates (CDR), breast MR (breast MRI) can reduce underdiagnosis of breast cancer compared to conventional imaging techniques, particularly in women with dense breasts. The purpose of this study is to report the additional breast cancer yield by breast MRI in women with dense breasts after receiving a negative screening mammogram. METHODS: For this study we invited consecutive participants of the national German breast cancer Screening program with breast density categories ACR C & D and a negative mammogram to undergo additional screening by breast MRI. Endpoints were CDR and recall rates. This study reports interim results in the first 200 patients. At a power of 80% and considering an alpha error of 5%, this preliminary population size is sufficient to demonstrate a 4/1000 improvement in CDR. RESULTS: In 200 screening participants, 8 women (40/1000, 17.4-77.3/1000) were recalled due to positive breast MRI findings. Image-guided biopsy revealed 5 cancers in 4 patients (one bilateral), comprising four invasive cancers and one case of DCIS. 3 patients revealed 4 invasive cancers presenting with ACR C breast density and one patient non-calcifying DCIS in a woman with ACR D breast density, resulting in a CDR of 20/1000 (95%-CI 5.5-50.4/1000) and a PPV of 50% (95%-CI 15.7-84.3%). CONCLUSION: Our initial results demonstrate that supplemental screening using breast MRI in women with heterogeneously dense and very dense breasts yields an additional cancer detection rate in line with a prior randomized trial on breast MRI screening of women with extremely dense breasts. These findings are highly important as the population investigated constitutes a much higher proportion of women and yielded cancers particularly in women with heterogeneously dense breasts.

Contrast-enhanced breast imaging: Current status and future challenges.

BACKGROUND: Contrast-enhanced breast MRI and recently also contrast-enhanced mammography (CEM) are available for breast imaging. The aim of the current overview is to explore existing evidence and ongoing challenges of contrast-enhanced breast imaging. METHODS: This narrative provides an introduction to the contrast-enhanced breast imaging modalities breast MRI and CEM. Underlying principle, techniques and BI-RADS reporting of both techniques are described and compared, and the following indications and ongoing challenges are discussed: problem-solving, high-risk screening, supplemental screening in women with extremely dense breast tissue, breast implants, neoadjuvant systemic therapy (NST) response monitoring, MRI-guided and CEM- guided biopsy. RESULTS: Technique and reporting for breast MRI are standardised, for the newer CEM standardisation is in progress. Similarly, compared to other modalities, breast MRI is well established as superior for problem-solving, screening women at high risk, screening women with extremely dense breast tissue or with implants; and for monitoring response to NST. Furthermore, MRI-guided biopsy is a reliable technique with low long-term false negative rates. For CEM, data is as yet either absent or limited, but existing results in these settings are promising. CONCLUSION: Contrast-enhanced breast imaging achieves highest diagnostic performance and should be considered essential. Of the two contrast-enhanced modalities, evidence of breast MRI superiority is ample, and preliminary results on CEM are promising, yet CEM warrants further study.

Using the Kaiser Score as a clinical decision rule for breast lesion classification: Does computer-assisted curve type analysis improve diagnosis?

PURPOSE: We aimed to investigate the effect of using visual or automatic enhancement curve type assessment on the diagnostic performance of the Kaiser Score (KS), a clinical decision rule for breast MRI. METHOD: This IRB-approved retrospective study analyzed consecutive conventional BI-RADS 0, 4 or 5 patients who underwent biopsy after 1.5T breast MRI according to EUSOBI recommendations between 2013 and 2015. The KS includes five criteria (spiculations; signal intensity (SI)-time curve type; margins of the lesion; internal enhancement; and presence of edema) resulting in scores from 1 (=lowest) to 11 (=highest risk of breast cancer). Enhancement curve types (Persistent, Plateau or Wash-out) were assessed by two radiologists independently visually and using a pixel-wise color-coded computed parametric map of curve types. KS diagnostic performance differences between readings were compared by ROC analysis. RESULTS: In total 220 lesions (147 benign, 73 malignant) including mass (n = 148) and non-mass lesions (n = 72) were analyzed. KS reading performance in distinguishing benign from malignant lesions did not differ between visual analysis and parametric map (P = 0.119; visual: AUC 0.875, sensitivity 95 %, specificity 63 %; and map: AUC 0.901, sensitivity 97 %, specificity 65 %). Additionally, analyzing mass and non-mass lesions separately, showed no difference between parametric map based and visual curve type-based KS analysis as well (P = 0.130 and P = 0.787). CONCLUSIONS: The performance of the Kaiser Score is largely independent of the curve type assessment methodology, confirming its robustness as a clinical decision rule for breast MRI in any type of breast lesion in clinical routine.

External validation of nomograms including MRI features for the prediction of side-specific extraprostatic extension.

BACKGROUND: Prediction of side-specific extraprostatic extension (EPE) is crucial in selecting patients for nerve-sparing radical prostatectomy (RP). Multiple nomograms, which include magnetic resonance imaging (MRI) information, are available predict side-specific EPE. It is crucial that the accuracy of these nomograms is assessed with external validation to ensure they can be used in clinical practice to support medical decision-making. METHODS: Data of prostate cancer (PCa) patients that underwent robot-assisted RP (RARP) from 2017 to 2021 at four European tertiary referral centers were collected retrospectively. Four previously developed nomograms for the prediction of side-specific EPE were identified and externally validated. Discrimination (area under the curve [AUC]), calibration and net benefit of four nomograms were assessed. To assess the strongest predictor among the MRI features included in all nomograms, we evaluated their association with side-specific EPE using multivariate regression analysis and Akaike Information Criterion (AIC). RESULTS: This study involved 773 patients with a total of 1546 prostate lobes. EPE was found in 338 (22%) lobes. The AUCs of the models predicting EPE ranged from 72.2% (95% CI 69.1-72.3%) (Wibmer) to 75.5% (95% CI 72.5-78.5%) (Nyarangi-Dix). The nomogram with the highest AUC varied across the cohorts. The Soeterik, Nyarangi-Dix, and Martini nomograms demonstrated fair to good calibration for clinically most relevant thresholds between 5 and 30%. In contrast, the Wibmer nomogram showed substantial overestimation of EPE risk for thresholds above 25%. The Nyarangi-Dix nomogram demonstrated a higher net benefit for risk thresholds between 20 and 30% when compared to the other three nomograms. Of all MRI features, the European Society of Urogenital Radiology score and tumor capsule contact length showed the highest AUCs and lowest AIC. CONCLUSION: The Nyarangi-Dix, Martini and Soeterik nomograms resulted in accurate EPE prediction and are therefore suitable to support medical decision-making.

Estimation of the biomechanical mammographic deformation of the breast using machine learning models.

BACKGROUND: A typical problem in the registration of MRI and X-ray mammography is the nonlinear deformation applied to the breast during mammography. We have developed a method for virtual deformation of the breast using a biomechanical model automatically constructed from MRI. The virtual deformation is applied in two steps: unloaded state estimation and compression simulation. The finite element method is used to solve the deformation process. However, the extensive computational cost prevents its usage in clinical routine. METHODS: We propose three machine learning models to overcome this problem: an extremely randomized tree (first model), extreme gradient boosting (second model), and deep learning-based bidirectional long short-term memory with an attention layer (third model) to predict the deformation of a biomechanical model. We evaluated our methods with 516 breasts with realistic compression ratios up to 76%. FINDINGS: We first applied one-fold validation, in which the second and third models performed better than the first model. We then applied ten-fold validation. For the unloaded state estimation, the median RMSE for the second and third models is 0.8 mm and 1.2 mm, respectively. For the compression, the median RMSE is 3.4 mm for both models. We evaluated correlations between model accuracy and characteristics of the clinical datasets such as compression ratio, breast volume, and tissue types. INTERPRETATION: Using the proposed models, we achieved accurate results comparable to the finite element model, with a speedup of factor 240 using the extreme gradient boosting model. These proposed models can replace the finite element model simulation, enabling clinically relevant real-time application.

How much can abbreviated protocols for breast MRI increase patient throughput? a multi-centric evaluation.

PURPOSE: To assess the impact of abbreviated breast MRI protocols on patient throughput considering non-scanning time and differences between in- and out-of-hospital settings. MATERIALS & METHODS: A total of 143 breast MRI exams from four study sites (hospital, three radiology centers) were included in this retrospective study. Total exam time (TET), Table Time (TT), Scan Time (ST), Table Switch Time (TST) and Planning Time (PT) were determined from consecutive breast MRI examinations. Possible number of scans and exams per hour were calculated. Four scan protocols were compared: full diagnostic protocol (n = 34, hospital), split dynamic protocol (n = 109, all sites) and two abbreviated protocols (n = 109, calculated, all sites). Data were described as median and interquartile range (IQR) and compared by Mann-Whitney-U-Test. RESULTS: Non-scanning time increased from 50% to 74% of the TET with a TST of 46% and a PT of 28% in the shortest abbreviated protocol. Number of possible scans per hour increased from 4.7 to 18.8 while number of possible exams per hour only increased from 2.3 to 5.1. Absolute TST (4.7 vs. 5.7 min, p = 0.46) and TET (18 min each, p = 0.35) did not differ significantly between in- and out-of-hospital exams. Absolute (4.4 vs. 2.8 min, p < 0.001) and relative (23 vs. 13%, p < 0.001) PT and TT (13.3 vs. 11.5 min, p = 0.004) was longer and relative TST (27% vs. 34%, p = 0.047) was shorter in hospital. CONCLUSION: TST and PT significantly contribute to TET and challenge the effectiveness of abbreviated protocols for increasing patient throughput. These findings show only low setting-dependent differences.

Impact of PI-QUAL on PI-RADS and cancer yield in an MRI-TRUS fusion biopsy population.

PURPOSE: To test the inter-reader agreement of the Prostate Imaging Quality (PI-QUAL) score for multiparametric prostate MRI and its impact on diagnostic performance in an MRI-ultrasound fusion biopsy population. PATIENTS AND METHODS: Pre-biopsy multiparametric (T2-weighted, DWI, and DCE) prostate MRIs (mpMRI) of 50 patients undergoing transrectal ultrasound-guided MRI-fusion (MRI-TRUS) biopsy were included. Two radiologists independently assigned a PI-QUAL score to each patient and assessed the diagnostic quality of individual sequences. PI-RADS categories were assigned to six regions per prostate (left and right: base/mid-glandular/apex). Inter-reader agreement was calculated using Cohen's kappa and diagnostic performance was compared by the area under the receiver operating characteristics curve (AUC). RESULTS: In 274 diagnostic areas, the malignancy rate was 62.7% (22.5% clinically significant prostate cancer, ISUP ≥ 2). Inter-reader agreement for the diagnostic quality was poor for T2w (kappa 0.19), fair for DWI and DCE (kappa 0.23 and 0.29) and moderate for PI-QUAL (kappa 0.51). For PI-RADS category assignments, inter-reader agreement was very good (kappa 0.86). Overall diagnostic performance did not differ between studies with a PI-QUAL score > 3 compared to a score ≤ 3 (p = 0.552; AUC 0.805 and 0.839). However, the prevalence of prostate cancer was significantly lower when the PI-QUAL score was ≤ 3 (16.7% vs. 30.2%, p = 0.008). CONCLUSION: PI-QUAL has only a limited impact on PI-RADS diagnostic performance in patients scheduled for MRI-TRUS fusion biopsy. However, the lower cancer prevalence in the lower PI-QUAL categories points out a risk of false-positive referrals and unnecessary biopsies if prostate imaging quality is low.

AI-enhanced simultaneous multiparametric 18F-FDG PET/MRI for accurate breast cancer diagnosis.

PURPOSE: To assess whether a radiomics and machine learning (ML) model combining quantitative parameters and radiomics features extracted from simultaneous multiparametric 18F-FDG PET/MRI can discriminate between benign and malignant breast lesions. METHODS: A population of 102 patients with 120 breast lesions (101 malignant and 19 benign) detected on ultrasound and/or mammography was prospectively enrolled. All patients underwent hybrid 18F-FDG PET/MRI for diagnostic purposes. Quantitative parameters were extracted from DCE (MTT, VD, PF), DW (mean ADC of breast lesions and contralateral breast parenchyma), PET (SUVmax, SUVmean, and SUVminimum of breast lesions, as well as SUVmean of the contralateral breast parenchyma), and T2-weighted images. Radiomics features were extracted from DCE, T2-weighted, ADC, and PET images. Different diagnostic models were developed using a fine Gaussian support vector machine algorithm which explored different combinations of quantitative parameters and radiomics features to obtain the highest accuracy in discriminating between benign and malignant breast lesions using fivefold cross-validation. The performance of the best radiomics and ML model was compared with that of expert reader review using McNemar's test. RESULTS: Eight radiomics models were developed. The integrated model combining MTT and ADC with radiomics features extracted from PET and ADC images obtained the highest accuracy for breast cancer diagnosis (AUC 0.983), although its accuracy was not significantly higher than that of expert reader review (AUC 0.868) (p = 0.508). CONCLUSION: A radiomics and ML model combining quantitative parameters and radiomics features extracted from simultaneous multiparametric 18F-FDG PET/MRI images can accurately discriminate between benign and malignant breast lesions.

Multiparametric ultrasound examination for response assessment in breast cancer patients undergoing neoadjuvant therapy.

To investigate the performance of multiparametric ultrasound for the evaluation of treatment response in breast cancer patients undergoing neoadjuvant chemotherapy (NAC). The IRB approved this prospective study. Breast cancer patients who were scheduled to undergo NAC were invited to participate in this study. Changes in tumour echogenicity, stiffness, maximum diameter, vascularity and integrated backscatter coefficient (IBC) were assessed prior to treatment and 7 days after four consecutive NAC cycles. Residual malignant cell (RMC) measurement at surgery was considered as standard of reference. RMC < 30% was considered a good response and > 70% a poor response. The correlation coefficients of these parameters were compared with RMC from post-operative histology. Linear Discriminant Analysis (LDA), cross-validation and Receiver Operating Characteristic curve (ROC) analysis were performed. Thirty patients (mean age 56.4 year) with 42 lesions were included. There was a significant correlation between RMC and echogenicity and tumour diameter after the 3rd course of NAC and average stiffness after the 2nd course. The correlation coefficient for IBC and echogenicity calculated after the first four doses of NAC were 0.27, 0.35, 0.41 and 0.30, respectively. Multivariate analysis of the echogenicity and stiffness after the third NAC revealed a sensitivity of 82%, specificity of 90%, PPV = 75%, NPV = 93%, accuracy = 88% and AUC of 0.88 for non-responding tumours (RMC > 70%). High tumour stiffness and persistent hypoechogenicity after the third NAC course allowed to accurately predict a group of non-responding tumours. A correlation between echogenicity and IBC was demonstrated as well.

Supervised machine learning enables non-invasive lesion characterization in primary prostate cancer with [68Ga]Ga-PSMA-11 PET/MRI.

PURPOSE: Risk classification of primary prostate cancer in clinical routine is mainly based on prostate-specific antigen (PSA) levels, Gleason scores from biopsy samples, and tumor-nodes-metastasis (TNM) staging. This study aimed to investigate the diagnostic performance of positron emission tomography/magnetic resonance imaging (PET/MRI) in vivo models for predicting low-vs-high lesion risk (LH) as well as biochemical recurrence (BCR) and overall patient risk (OPR) with machine learning. METHODS: Fifty-two patients who underwent multi-parametric dual-tracer [18F]FMC and [68Ga]Ga-PSMA-11 PET/MRI as well as radical prostatectomy between 2014 and 2015 were included as part of a single-center pilot to a randomized prospective trial (NCT02659527). Radiomics in combination with ensemble machine learning was applied including the [68Ga]Ga-PSMA-11 PET, the apparent diffusion coefficient, and the transverse relaxation time-weighted MRI scans of each patient to establish a low-vs-high risk lesion prediction model (MLH). Furthermore, MBCR and MOPR predictive model schemes were built by combining MLH, PSA, and clinical stage values of patients. Performance evaluation of the established models was performed with 1000-fold Monte Carlo (MC) cross-validation. Results were additionally compared to conventional [68Ga]Ga-PSMA-11 standardized uptake value (SUV) analyses. RESULTS: The area under the receiver operator characteristic curve (AUC) of the MLH model (0.86) was higher than the AUC of the [68Ga]Ga-PSMA-11 SUVmax analysis (0.80). MC cross-validation revealed 89% and 91% accuracies with 0.90 and 0.94 AUCs for the MBCR and MOPR models respectively, while standard routine analysis based on PSA, biopsy Gleason score, and TNM staging resulted in 69% and 70% accuracies to predict BCR and OPR respectively. CONCLUSION: Our results demonstrate the potential to enhance risk classification in primary prostate cancer patients built on PET/MRI radiomics and machine learning without biopsy sampling.

Deep learning analysis provides accurate COVID-19 diagnosis on chest computed tomography.

INTRODUCTION: Computed Tomography is an essential diagnostic tool in the management of COVID-19. Considering the large amount of examinations in high case-load scenarios, an automated tool could facilitate and save critical time in the diagnosis and risk stratification of the disease. METHODS: A novel deep learning derived machine learning (ML) classifier was developed using a simplified programming approach and an open source dataset consisting of 6868 chest CT images from 418 patients which was split into training and validation subsets. The diagnostic performance was then evaluated and compared to experienced radiologists on an independent testing dataset. Diagnostic performance metrics were calculated using Receiver Operating Characteristics (ROC) analysis. Operating points with high positive (>10) and low negative (<0.01) likelihood ratios to stratify the risk of COVID-19 being present were identified and validated. RESULTS: The model achieved an overall accuracy of 0.956 (AUC) on an independent testing dataset of 90 patients. Both rule-in and rule out thresholds were identified and tested. At the rule-in operating point, sensitivity and specificity were 84.4 % and 93.3 % and did not differ from both radiologists (p > 0.05). At the rule-out threshold, sensitivity (100 %) and specificity (60 %) differed significantly from the radiologists (p < 0.05). Likelihood ratios and a Fagan nomogram provide prevalence independent test performance estimates. CONCLUSION: Accurate diagnosis of COVID-19 using a basic deep learning approach is feasible using open-source CT image data. In addition, the machine learning classifier provided validated rule-in and rule-out criteria could be used to stratify the risk of COVID-19 being present.

Combined texture analysis and machine learning in suspicious calcifications detected by mammography: Potential to avoid unnecessary stereotactical biopsies.

OBJECTIVES: To investigate whether combined texture analysis and machine learning can distinguish malignant from benign suspicious mammographic calcifications, to find an exploratory rule-out criterion to potentially avoid unnecessary benign biopsies. METHODS: Magnification views of 235 patients which underwent vacuum-assisted biopsy of suspicious calcifications (BI-RADS 4) during a two-year period were retrospectively analyzed using the texture analysis tool MaZda (Version 4.6). Microcalcifications were manually segmented and analyzed by two readers, resulting in 249 image features from gray-value histogram, gray-level co-occurrence and run-length matrices. After feature reduction with principal component analysis (PCA), a multilayer perceptron (MLP) artificial neural network was trained using histological results as the reference standard. For training and testing of this model, the dataset was split into 70 % and 30 %. ROC analysis was used to calculate diagnostic performance indices. RESULTS: 226 patients (150 benign, 76 malignant) were included in the final analysis due to missing data in 9 cases. Feature selection yielded nine image features for MLP training. Area under the ROC-curve in the testing dataset (n = 54) was 0.82 (95 %-CI: 0.70-0.94) and 0.832 (95 %-CI 0.72-0.94) for both readers, respectively. A high sensitivity threshold criterion was identified in the training dataset and successfully applied to the testing dataset, demonstrating the potential to avoid 37.1-45.7 % of unnecessary biopsies at the cost of one false-negative for each reader. CONCLUSION: Combined texture analysis and machine learning could be used for risk stratification in suspicious mammographic calcifications. At low costs in terms of false-negatives, unnecessary biopsies could be avoided.

Diagnostic performance of breast tumor tissue selection in diffusion weighted imaging: A systematic review and meta-analysis.

BACKGROUND: Several methods for tumor delineation are used in literature on breast diffusion weighted imaging (DWI) to measure the apparent diffusion coefficient (ADC). However, in the process of reaching consensus on breast DWI scanning protocol, image analysis and interpretation, still no standardized optimal breast tumor tissue selection (BTTS) method exists. Therefore, the purpose of this study is to assess the impact of BTTS methods on ADC in the discrimination of benign from malignant breast lesions in DWI in terms of sensitivity, specificity and area under the curve (AUC). METHODS AND FINDINGS: In this systematic review and meta-analysis, adhering to the PRISMA statement, 61 studies, with 65 study subsets, in females with benign or malignant primary breast lesions (6291 lesions) were assessed. Studies on DWI, quantified by ADC, scanned on 1.5 and 3.0 Tesla and using b-values 0/50 and ≥ 800 s/mm2 were included. PubMed and EMBASE were searched for studies up to 23-10-2019 (n = 2897). Data were pooled based on four BTTS methods (by definition of measured region of interest, ROI): BTTS1: whole breast tumor tissue selection, BTTS2: subtracted whole breast tumor tissue selection, BTTS3: circular breast tumor tissue selection and BTTS4: lowest diffusion breast tumor tissue selection. BTTS methods 2 and 3 excluded necrotic, cystic and hemorrhagic areas. Pooled sensitivity, specificity and AUC of the BTTS methods were calculated. Heterogeneity was explored using the inconsistency index (I2) and considering covariables: field strength, lowest b-value, image of BTTS selection, pre-or post-contrast DWI, slice thickness and ADC threshold. Pooled sensitivity, specificity and AUC were: 0.82 (0.72-0.89), 0.79 (0.65-0.89), 0.88 (0.85-0.90) for BTTS1; 0.91 (0.89-0.93), 0.84 (0.80-0.87), 0.94 (0.91-0.96) for BTTS2; 0.89 (0.86-0.92), 0.90 (0.85-0.93), 0.95 (0.93-0.96) for BTTS3 and 0.90 (0.86-0.93), 0.84 (0.81-0.87), 0.86 (0.82-0.88) for BTTS4, respectively. Significant heterogeneity was found between studies (I2 = 95). CONCLUSIONS: None of the breast tissue selection (BTTS) methodologies outperformed in differentiating benign from malignant breast lesions. The high heterogeneity of ADC data acquisition demands further standardization, such as DWI acquisition parameters and tumor tissue selection to substantially increase the reliability of DWI of the breast.

[Artificial intelligence in the diagnosis of breast cancer : Yesterday, today and tomorrow]. / Künstliche Intelligenz in der Mammadiagnostik : Gestern, heute und morgen.

BACKGROUND: Artificial intelligence (AI) is increasingly applied in the field of breast imaging. OBJECTIVES: What are the main areas where AI is applied in breast imaging and what AI and computer-aided diagnosis (CAD) systems are already available? MATERIALS AND METHODS: Basic literature and vendor-supplied information are screened for relevant information, which is then pooled, structured and discussed from the perspective of breast imaging. RESULTS: Original CAD systems in mammography date almost 25 years back. They are much more widely applied in the United States than in Europe. The initial CAD systems exhibited limited diagnostic abilities and disproportionally high rates of false positive results. Since 2012, deep learning mechanisms have been applied and expand the application possibilities of AI. CONCLUSION: To date there is no algorithm that has beyond doubt been proven to outperform double reporting by two certified breast radiologists. AI could, however, in the foreseeable future, take over the following tasks: preselection of abnormal examinations to substantially reduce workload of the radiologists by either excluding normal findings from human review or by replacing the double reader in screening. Furthermore, the establishment of radio-patho-genomic correlations and their translation into clinical practice is hardly conceivable without AI.

Impact of the Kaiser score on clinical decision-making in BI-RADS 4 mammographic calcifications examined with breast MRI.

OBJECTIVES: To investigate whether the application of the Kaiser score for breast magnetic resonance imaging (MRI) might downgrade breast lesions that present as mammographic calcifications and avoid unnecessary breast biopsies METHODS: This IRB-approved, retrospective, cross-sectional, single-center study included 167 consecutive patients with suspicious mammographic calcifications and histopathologically verified results. These patients underwent a pre-interventional breast MRI exam for further diagnostic assessment before vacuum-assisted stereotactic-guided biopsy (95 malignant and 72 benign lesions). Two breast radiologists with different levels of experience independently read all examinations using the Kaiser score, a machine learning-derived clinical decision-making tool that provides probabilities of malignancy by a formalized combination of diagnostic criteria. Diagnostic performance was assessed by receiver operating characteristics (ROC) analysis and inter-reader agreement by the calculation of Cohen's kappa coefficients. RESULTS: Application of the Kaiser score revealed a large area under the ROC curve (0.859-0.889). Rule-out criteria, with high sensitivity, were applied to mass and non-mass lesions alike. The rate of potentially avoidable breast biopsies ranged between 58.3 and 65.3%, with the lowest rate observed with the least experienced reader. CONCLUSIONS: Applying the Kaiser score to breast MRI allows stratifying the risk of breast cancer in lesions that present as suspicious calcifications on mammography and may thus avoid unnecessary breast biopsies. KEY POINTS: • The Kaiser score is a helpful clinical decision tool for distinguishing malignant from benign breast lesions that present as calcifications on mammography. • Application of the Kaiser score may obviate 58.3-65.3% of unnecessary stereotactic biopsies of suspicious calcifications. • High Kaiser scores predict breast cancer with high specificity, aiding clinical decision-making with regard to re-biopsy in case of negative results.

Can the addition of clinical information improve the accuracy of PI-RADS version 2 for the diagnosis of clinically significant prostate cancer in positive MRI?

AIM: To report prostate cancer (PCa) prevalence in Prostate Imaging Reporting and Data System version 2 (PI-RADS v2) categories and investigate the potential to avoid unnecessary, magnetic resonance imaging (MRI)-guided in-bore biopsies by adding clinical and biochemical patient characteristics. MATERIALS AND METHODS: The present institutional review board-approved, prospective study on 137 consecutive men with 178 suspicious lesions on 3 T MRI was performed. Routine data collected for each patient included patient characteristics (age, prostate volume), clinical background information (prostate-specific antigen [PSA] levels, PSA density), and PI-RADS v2 scores assigned in a double-reading approach. RESULTS: Histopathological evaluation revealed a total of 93/178 PCa (52.2%). The mean age was 66.3 years and PSA density was 0.24 ng/ml2 (range, 0.04-0.89 ng/ml). Clinically significant PCa (csPCa, Gleason score >6) was confirmed in 50/93 (53.8%) lesions and was significantly associated with higher PI-RADS v2 scores (p=0.0044). On logistic regression analyses, age, PSA density, and PI-RADS v2 scores contributed independently to the diagnosis of csPCa (p=7.9×10-7, p=0.097, and p=0.024, respectively). The resulting area under the receiver operating characteristic curve (AUC) to predict csPCa was 0.76 for PI-RADS v2, 0.59 for age, and 0.67 for PSA density. The combined regression model yielded an AUC of 0.84 for the diagnosis of csPCa and was significantly superior to each single parameter (p≤0.0009, respectively). Unnecessary biopsies could have been avoided in 50% (64/128) while only 4% (2/50) of csPCa lesions would have been missed. CONCLUSIONS: Adding age and PSA density to PI-RADS v2 scores improves the diagnostic accuracy for csPCa. A combination of these variables with PI-RADS v2 can help to avoid unnecessary in-bore biopsies while still detecting the majority of csPCa.

[Contrast-agent free imaging of the prostate : Ready for clinical practice?] / Kontrastmittelfreie Bildgebung der Prostata : Reif für die Praxis?

BACKGROUND: Multiparametric MRI (mpMRI) is currently the most accurate imaging modality for detection and local staging of prostate cancer (PCa). Disadvantages of this modality are high costs, time consumption and the need for a contrast medium. AIMS: The aim of the work was to provide an overview of the current state of fast and contrast-free MRI imaging of the prostate. RESULTS: Biparametric examination protocols and the use of three-dimensional T2-weighted sequences are readily available methods that can be used to shorten the examination time without sacrificing diagnostic accuracy.

[Contrast agent-free breast MRI : Advantages and potential disadvantages]. / Kontrastmittelfreie Mamma-MRT : Vorteile und potenzielle Nachteile.

BACKGROUND: Contrast-enhanced breast magnetic resonance imaging (MRI) is the most sensitive method for detection of breast cancer. The further spread of breast MRI is limited by the complicated examination procedure and the need for intravenously administered contrast media. OBJECTIVES: Can diffusion-weighted imaging (DWI) replace contrast-enhanced sequences to achieve an unenhanced breast MRI examination? MATERIALS AND METHODS: Narrative review and meta-analytic assessment of previously published studies. RESULTS: DWI can visualize breast lesions and distinguish benign from malignant findings. It is thus a valid alternative to contrast-enhanced sequences. As an additional technique, the use of DWI can reduce the numbers of unnecessary breast biopsies. The lack of robustness leading to variable sensitivity that is currently lower than that of contrast-enhanced breast MRI is a disadvantage of DWI. CONCLUSIONS: Presently, DWI can be recommended as an integral part of clinical breast MRI protocols. The application as a stand-alone technique within unenhanced protocols is still under evaluation.

Impact of androgen deprivation therapy on apparent diffusion coefficient and T2w MRI for histogram and texture analysis with respect to focal radiotherapy of prostate cancer.

PURPOSE: Accurate prostate cancer (PCa) detection is essential for planning focal external beam radiotherapy (EBRT). While biparametric MRI (bpMRI) including T2-weighted (T2w) and diffusion-weighted images (DWI) is an accurate tool to localize PCa, its value is less clear in the case of additional androgen deprivation therapy (ADT). The aim of this study was to investigate the value of a textural feature (TF) approach on bpMRI analysis in prostate cancer patients with and without neoadjuvant ADT with respect to future dose-painting applications. METHODS: 28 PCa patients (54-80 years) with (n = 14) and without (n = 14) ADT who underwent bpMRI with T2w and DWI were analyzed retrospectively. Lesions, central gland (CG), and peripheral zone (PZ) were delineated by an experienced urogenital radiologist based on localized pre-therapeutic histopathology. Histogram parameters and 20 Haralick TF were calculated. Regional differences (i. e., tumor vs. PZ, tumor vs. CG) were analyzed for all imaging parameters. Receiver-operating characteristic (ROC) analysis was performed to measure diagnostic performance to distinguish PCa from benign prostate tissue and to identify the features with best discriminative power in both patient groups. RESULTS: The obtained sensitivities were equivalent or superior when utilizing the TF in the no-ADT group, while specificity was higher for the histogram parameters. However, in the ADT group, TF outperformed the conventional histogram parameters in both specificity and sensitivity. Rule-in and rule-out criteria for ADT patients could exclusively be defined with the aid of TF. CONCLUSIONS: The TF approach has the potential for quantitative image-assisted boost volume delineation in PCa patients even if they are undergoing neoadjuvant ADT.