Investigating the prediction value of multiparametric magnetic resonance imaging at 3 T in response to neoadjuvant chemotherapy in breast cancer

Longitudinal ultrasound-based AI model predicts axillary lymph node response to neoadjuvant chemotherapy in breast cancer: a multicenter study

OBJECTIVES

Developing a deep learning radiomics model from longitudinal breast ultrasound and sonographer's axillary ultrasound diagnosis for predicting axillary lymph node (ALN) response to neoadjuvant chemotherapy (NAC) in breast cancer.

METHODS

Breast cancer patients undergoing NAC followed by surgery were recruited from three centers between November 2016 and December 2022. We collected ultrasound images for extracting tumor-derived radiomics and deep learning features, selecting quantitative features through various methods. Two machine learning models based on random forest were developed using pre-NAC and post-NAC features. A support vector machine integrated these data into a fusion model, evaluated via the area under the curve (AUC), decision curve analysis, and calibration curves. We compared the fusion model's performance against sonographer's diagnosis from pre-NAC and post-NAC axillary ultrasonography, referencing histological outcomes from sentinel lymph node biopsy or axillary lymph node dissection.

RESULTS

In the validation cohort, the fusion model outperformed both pre-NAC (AUC: 0.899 vs. 0.786, p < 0.001) and post-NAC models (AUC: 0.899 vs. 0.853, p = 0.014), as well as the sonographer's diagnosis of ALN status on pre-NAC and post-NAC axillary ultrasonography (AUC: 0.899 vs. 0.719, p < 0.001). Decision curve analysis revealed patient benefits from the fusion model across threshold probabilities from 0.02 to 0.98. The model also enhanced sonographer's diagnostic ability, increasing accuracy from 71.9% to 79.2%.

CONCLUSION

The deep learning radiomics model accurately predicted the ALN response to NAC in breast cancer. Furthermore, the model will assist sonographers to improve their diagnostic ability on ALN status before surgery.

CLINICAL RELEVANCE STATEMENT

Our AI model based on pre- and post-neoadjuvant chemotherapy ultrasound can accurately predict axillary lymph node metastasis and assist sonographer's axillary diagnosis.

KEY POINTS

Axillary lymph node metastasis status affects the choice of surgical treatment, and currently relies on subjective ultrasound. Our AI model outperformed sonographer's visual diagnosis on axillary ultrasound. Our deep learning radiomics model can improve sonographers' diagnosis and might assist in surgical decision-making.

A non-invasive artificial intelligence model for identifying axillary pathological complete response to neoadjuvant chemotherapy in breast cancer: a secondary analysis to multicenter clinical trial

BACKGROUND

This study aims to develop a stacking model for accurately predicting axillary lymph node (ALN) response to neoadjuvant chemotherapy (NAC) using longitudinal MRI in breast cancer.

METHODS

We included patients with node-positive breast cancer who received NAC following surgery from January 2012 to June 2022. We collected MRIs before and after NAC, and extracted radiomics features from the tumour, peritumour, and ALN regions. The Mann-Whitney U test, least absolute shrinkage and selection operator, and Boruta algorithm were used to select features. We utilised machine learning techniques to develop three single-modality models and a stacking model for predicting ALN response to NAC.

RESULTS

This study consisted of a training cohort (n = 277), three external validation cohorts (n = 313, 164, and 318), and a prospective cohort (n = 81). Among the 1153 patients, 60.62% achieved ypN0. The stacking model achieved excellent AUCs of 0.926, 0.874, and 0.862 in the training, external validation, and prospective cohort, respectively. It also showed lower false-negative rates (FNRs) compared to radiologists, with rates of 14.40%, 20.85%, and 18.18% (radiologists: 40.80%, 50.49%, and 63.64%) in three cohorts. Additionally, there was a significant difference in disease-free survival between high-risk and low-risk groups (p < 0.05).

CONCLUSIONS

The stacking model can accurately predict ALN status after NAC in breast cancer, showing a lower false-negative rate than radiologists.

TRIAL REGISTRATION NUMBER

The clinical trial numbers were NCT03154749 and NCT04858529.

The longitudinal changes in multiparametric MRI during neoadjuvant chemotherapy can predict treatment response early in patients with HER2-positive breast cancer

PURPOSE

To investigate whether longitudinal changes in multiparametric MRI can predict early response to neoadjuvant chemotherapy (NAC) for HER2-positive breast cancer (BC) and to further establish quantitative models based on these features.

METHODS

A total of 164 HER2-positive BC patients from three centers were included. MRI was performed at baseline and after two cycles of NAC (early post-NAC). Clinicopathological characteristics were enrolled. MRI features were evaluated at baseline and early post-NAC, as well as longitudinal changes in multiparametric MRI, including changes in the largest diameter (LD) of the tumor (ΔLD), apparent diffusion coefficient (ADC) values (ΔADC), and time-signal intensity curve (TIC) (ΔTIC). The patients were divided into a training set (n = 95), an internal validation set (n = 31), and an independent external validation set (n = 38). Univariate and multivariate logistic regression analyses were used to identify the independent indicators of pCR, which were then used to establish the clinicopathologic model and combined model. The AUC was used to evaluate the predictive power of the different models and calibration curves were used to evaluate the consistency of the prediction of pCR in different models. Additionally, decision curve analysis (DCA) was employed to determine the clinical usefulness of the different models.

RESULTS

Two models were enrolled in this study, including the clinicopathologic model and the combined model. The LD at early post-NAC (OR=0.913, 95 % CI=0.953-0.994 p = 0.026), ΔADC (OR=1.005, 95 % CI=1.005-1.008, p = 0.007), and ΔTIC (OR=3.974, 95 % CI=1.276-12.358, p = 0.017) were identified as the best predictors of NAC response. The combined model constructed by the combination of LD at early post-NAC, ΔADC, and ΔTIC showed good predictive performance in the training set (AUC=0.87), internal validation set (AUC=0.78), and external validation set (AUC=0.79), which performed better than the clinicopathologic model in all sets.

CONCLUSIONS

The changes in multiparametric MRI can predict early treatment response for HER2-positive BC and may be helpful for individualized treatment planning.

Development and Validation of an Interpretable Machine Learning Prediction Model for Total Pathological Complete Response after Neoadjuvant Chemotherapy in Locally Advanced Breast Cancer: Multicenter Retrospective Analysis

Objective: This study aims to develop an interpretable machine learning (ML) model to accurately predict the probability of achieving total pathological complete response (tpCR) in patients with locally advanced breast cancer (LABC) following neoadjuvant chemotherapy (NAC). Methods: This multi-center retrospective study included pre-NAC clinical pathology data from 698 LABC patients. Post-operative pathological outcomes divided patients into tpCR and non-tpCR groups. Data from 586 patients at Shanghai Ruijin Hospital were randomly assigned to a training set (80%) and a test set (20%). In comparison, data from our hospital's remaining 112 patients were used for external validation. Variable selection was performed using the Least Absolute Shrinkage and Selection Operator (LASSO) regression analysis. Predictive models were constructed using six ML algorithms: decision trees, K-nearest neighbors (KNN), support vector machine, light gradient boosting machine, and extreme gradient boosting. Model efficacy was assessed through various metrics, including receiver operating characteristic (ROC) curves, precision-recall (PR) curves, confusion matrices, calibration plots, and decision curve analysis (DCA). The best-performing model was selected by comparing the performance of different algorithms. Moreover, variable relevance was ranked using the SHapley Additive exPlanations (SHAP) technique to improve the interpretability of the model and solve the "black box" problem. Results: A total of 191 patients (32.59%) achieved tpCR following NAC. Through LASSO regression analysis, five variables were identified as predictive factors for model construction, including tumor size, Ki-67, molecular subtype, targeted therapy, and chemotherapy regimen. The KNN model outperformed the other five classifier algorithms, achieving area under the curve (AUC) values of 0.847 (95% CI: 0.809-0.883) in the training set, 0.763 (95% CI: 0.670-0.856) in the test set, and 0.665 (95% CI: 0.555-0.776) in the external validation set. DCA demonstrated that the KNN model yielded the highest net advantage through a wide range of threshold probabilities in both the training and test sets. Furthermore, the analysis of the KNN model utilizing SHAP technology demonstrated that targeted therapy is the most crucial factor in predicting tpCR. Conclusion: An ML prediction model using clinical and pathological data collected before NAC was developed and verified. This model accurately predicted the probability of achieving a tpCR in patients with LABC after receiving NAC. SHAP technology enhanced the interpretability of the model and assisted in clinical decision-making and therapy optimization.

Changes in Intratumor Blood Flow After Carbon-Ion Radiation Therapy for Early-Stage Breast Cancer

Purpose

This study aimed to quantify the changes in intratumoral blood flow after carbon-ion radiation therapy (CIRT) for early-stage breast cancer and analyze their clinical significance.

Patients and Methods

We included 38 patients with early-stage breast cancer who underwent CIRT. Dynamic imaging was performed using a 3T superconducting magnetic resonance scanner to quantify the washin index (idx), which reflects contrast uptake, and washout idx, which reflects the rate of contrast washout from tumor tissue. The changes in the apparent diffusion coefficient, washin idx, and washout idx were examined before CIRT and at 1 and 3 months after treatment. Clinical factors and imaging features were examined using univariate and receiver operating characteristic curve analyses to identify factors predicting clinical complete response (cCR).

Results

The median observation period after CIRT was 51 (range: 12-122) months. During the observation period, 31 of the 38 patients achieved cCR, and 22 achieved cCR within 12 months. Tumor size (P < .001), washin idx (P = .043), and washout idx (P < .001) decreased significantly 1-month after CIRT. In contrast, the apparent diffusion coefficient values (P < .001) increased significantly 1-month after CIRT. Univariate analysis suggested that the washin idx after 1 and 3 months of CIRT was associated with cCR by 12 months post-CIRT (P = .028 and .021, respectively). No other parameters were associated with cCR by 12 months post-CIRT. Furthermore, receiver operating characteristic curve analyses showed that the area under the curve values of washin idx after 1 and 3 months of CIRT was 0.78 (specificity 75%, sensitivity 80%) and 0.73 (specificity 75%, sensitivity 71%), respectively.

Conclusion

Tumor changes can be quantified early after CIRT using contrast-enhanced magnetic resonance imaging in patients with breast cancer. Washin idx values 1 and 3 months after CIRT were associated with cCR within 12 months post-CIRT.

Early prediction of pathological response to neoadjuvant chemotherapy of breast tumors: a comparative study using amide proton transfer-weighted, diffusion weighted and dynamic contrast enhanced MRI

Objective

To examine amide proton transfer-weighted (APTw) combined with diffusion weighed (DWI) and dynamic contrast enhanced (DCE) MRI for early prediction of pathological response to neoadjuvant chemotherapy in invasive breast cancer.

Materials

In this prospective study, 50 female breast cancer patients (49.58 ± 10.62 years old) administered neoadjuvant chemotherapy (NAC) were enrolled with MRI carried out both before NAC (T0) and at the end of the second cycle of NAC (T1). The patients were divided into 2 groups based on tumor response according to the Miller-Payne Grading (MPG) system. Group 1 included patients with a greater degree of decrease in major histologic responder (MHR, Miller-Payne G4-5), while group 2 included non-MHR cases (Miller-Payne G1-3). Traditional imaging protocols (T1 weighted, T2 weighted, diffusion weighted, and DCE-MRI) and APTw imaging were scanned for each subject before and after treatment. APTw value (APTw0 and APTw1), Dmax (maximum diameter, Dmax0 and Dmax1), V (3D tumor volume, V0 and V1), and ADC (apparent diffusion coefficient, ADC0 and ADC1) before and after treatment, as well as changes between the two times points (ΔAPT, ΔDmax, ΔV, ΔADC) for breast tumors were compared between the two groups.

Results

APT0 and APT1 values significantly differed between the two groups (p = 0.034 and 0.01). ΔAPTw values were significantly lower in non-MHR tumors compared with MHR tumors (p = 0.015). ΔDmax values were significantly higher in MHR tumors compared with non-MHR tumors (p = 0.005). ADC0 and ADC1 values were significantly higher in MHR tumors than in non-MHR tumors (p = 0.038 and 0.035). AUC (Dmax+DWI + APTw) = AUC (Dmax+APTw) > AUC (APTw) > AUC (Dmax+DWI) > AUC (Dmax).

Conclusion

APTw imaging along with change of tumor size showed a significant potential in early prediction of MHR for NAC treatment in breast cancer, which might allow timely regimen refinement before definitive surgical treatment.

Early prediction of pathologic complete response of breast cancer after neoadjuvant chemotherapy using longitudinal ultrafast dynamic contrast-enhanced MRI

PURPOSE

The purpose of this study was to evaluate the temporal trends of ultrafast dynamic contrast-enhanced (DCE)-MRI during neoadjuvant chemotherapy (NAC) and to investigate whether the changes in DCE-MRI parameters could early predict pathologic complete response (pCR) of breast cancer.

MATERIALS AND METHODS

This longitudinal study prospectively recruited consecutive participants with breast cancer who underwent ultrafast DCE-MRI examinations before treatment and after two, four, and six NAC cycles between February 2021 and February 2022. Five ultrafast DCE-MRI parameters (maximum slope [MS], time-to-peak [TTP], time-to-enhancement [TTE], peak enhancement intensity [PEI], and initial area under the curve in 60 s [iAUC]) and tumor size were measured at each timepoint. The changes in parameters between each pair of adjacent timepoints were additionally measured and compared between the pCR and non-pCR groups. Longitudinal data were analyzed using generalized estimating equations. The performance for predicting pCR was assessed using area under the receiver operating characteristic curve (AUC).

RESULTS

Sixty-seven women (mean age, 50 ± 8 [standard deviation] years; age range: 25-69 years) were included, 19 of whom achieved pCR. MS, PEI, iAUC, and tumor size decreased, while TTP increased during NAC (all P < 0.001). The AUC (0.92; 95% confidence interval [CI]: 0.83-0.97) of the model incorporating ultrafast DCE-MRI parameter change values (from timepoints 1 to 2) and clinicopathologic characteristics was greater than that of the clinical model (AUC, 0.79; 95% CI: 0.68-0.88) and ultrafast DCE-MRI parameter model at timepoint 2 when combined with clinicopathologic characteristics (AUC, 0.82; 95% CI: 0.71-0.90) (P = 0.01 and 0.02).

CONCLUSION

Early changes in ultrafast DCE-MRI parameters after NAC combined with clinicopathologic characteristics could serve as predictive markers of pCR of breast cancer.

Multifactor artificial intelligence model assists axillary lymph node surgery in breast cancer after neoadjuvant chemotherapy: multicenter retrospective cohort study

BACKGROUND

The high false negative rate (FNR) associated with sentinel lymph node biopsy often leads to unnecessary axillary lymph node dissection following neoadjuvant chemotherapy (NAC) in breast cancer. The authors aimed to develop a multifactor artificial intelligence (AI) model to aid in axillary lymph node surgery.

MATERIALS AND METHODS

A total of 1038 patients were enrolled, comprising 234 patients in the primary cohort, 723 patients in three external validation cohorts, and 81 patients in the prospective cohort. For predicting axillary lymph node response to NAC, robust longitudinal radiomics features were extracted from pre-NAC and post-NAC magnetic resonance images. The U test, the least absolute shrinkage and selection operator, and the spearman analysis were used to select the most significant features. A machine learning stacking model was constructed to detect ALN metastasis after NAC. By integrating the significant predictors, we developed a multifactor AI-assisted surgery pipeline and compared its performance and false negative rate with that of sentinel lymph node biopsy alone.

RESULTS

The machine learning stacking model achieved excellent performance in detecting ALN metastasis, with an area under the curve (AUC) of 0.958 in the primary cohort, 0.881 in the external validation cohorts, and 0.882 in the prospective cohort. Furthermore, the introduction of AI-assisted surgery reduced the FNRs from 14.88 (18/121) to 4.13% (5/121) in the primary cohort, from 16.55 (49/296) to 4.05% (12/296) in the external validation cohorts, and from 13.64 (3/22) to 4.55% (1/22) in the prospective cohort. Notably, when more than two SLNs were removed, the FNRs further decreased to 2.78% (2/72) in the primary cohort, 2.38% (4/168) in the external validation cohorts, and 0% (0/15) in the prospective cohort.

CONCLUSION

Our study highlights the potential of AI-assisted surgery as a valuable tool for evaluating ALN response to NAC, leading to a reduction in unnecessary axillary lymph node dissection procedures.

Correlation between Baseline Conventional Ultrasounds, Shear-Wave Elastography Indicators, and Neoadjuvant Therapy Efficacy in Triple-Negative Breast Cancer

In patients with triple-negative breast cancer (TNBC)-the subtype with the poorest prognosis among breast cancers-it is crucial to assess the response to the currently widely employed neoadjuvant treatment (NAT) approaches. This study investigates the correlation between baseline conventional ultrasound (US) and shear-wave elastography (SWE) indicators and the pathological response of TNBC following NAT, with a specific focus on assessing predictive capability in the baseline state. This retrospective analysis was conducted by extracting baseline US features and SWE parameters, categorizing patients based on postoperative pathological grading. A univariate analysis was employed to determine the relationship between ultrasound indicators and pathological reactions. Additionally, we employed a receiver operating characteristic (ROC) curve analysis and multivariate logistic regression methods to evaluate the predictive potential of the baseline US indicators. This study comprised 106 TNBC patients, with 30 (28.30%) in a nonmajor histological response (NMHR) group and 76 (71.70%) in a major histological response (MHR) group. Following the univariate analysis, we found that T staging, dmax values, volumes, margin changes, skin alterations (i.e., thickening and invasion), retromammary space invasions, and supraclavicular lymph node abnormalities were significantly associated with pathological efficacy (p < 0.05). Combining clinical information with either US or SWE independently yielded baseline predictive abilities, with AUCs of 0.816 and 0.734, respectively. Notably, the combined model demonstrated an improved AUC of 0.827, with an accuracy of 76.41%, a sensitivity of 90.47%, a specificity of 55.81%, and statistical significance (p < 0.01). The baseline US and SWE indicators for TNBC exhibited a strong relationship with NAT response, offering predictive insights before treatment initiation, to a considerable extent.

Radiomics Based on DCE-MRI for Predicting Response to Neoadjuvant Therapy in Breast Cancer

RATIONALE AND OBJECTIVES

To compare the value of radiomics and diameter% based on pre- and early-treatment dynamic enhanced MR (DCE-MRI) of the breast in predicting response to neoadjuvant therapy (NAT) in breast cancer and to construct a tool for early noninvasive prediction of NAT outcomes.

MATERIALS AND METHODS

Retrospective analysis of clinical and imaging data of 142 patients with primary invasive breast cancer who underwent DCE-MRI before and after two cycles of NAT at our institution. Enroled patients were randomly assigned in a 7:3 ratio to the training group and the test group. Patients were divided into pathological complete response (pCR) and non-pathological complete response groups based on surgical pathology findings after NAT. The maximum diameter relative regression values (Diameter%) before and after treatment were calculated and the conventional imaging Diameter% model was constructed. Based on pre- and early-NAT DCE-MRI, the optimal features of pre-NAT, early-NAT, and delta radiomics were screened using redundancy analysis, least absolute shrinkage, and selection operator methods to construct the corresponding radiomics model and calculate the Radscores. Indicators that were statistically significant in the univariate analysis of clinical data were further screened by stepwise regression and combined with Radscores to construct the fusion model. All models were evaluated and compared.

RESULTS

In the test set, the area under the curve (AUC) of the delta radiomics model (0.87) was higher than that of the pre-NAT, early-NAT radiomics models (0.57, 0.78) and the Diameter% model (0.83). The fusion model had the best efficacy in predicting pCR after NAT, with AUCs of 0.91 in the training and test sets. And its nomogram plot showed that Radscore of early-NAT radiomics had the greatest weight. In the test set, the fusion model and Delta radiomics model improved the efficacy of predicting pCR by 35.56% and 14.19%, respectively, compared to the Diameter% model (P = 0 and .039). Clinical decision curves showed the highest overall clinical benefit for the fusion model.

CONCLUSION

Radiomics, especially delta and early-NAT radiomics, may be potential biomarkers for early noninvasive prediction of NAT outcomes. And a fusion model constructed from meaningful clinicopathological indicators combined with radiomics can effectively predict NAT response.

Can peritumoral edema evaluated by Magnetic Resonance Imaging before neoadjuvant chemotherapy predict complete pathological response in breast cancer?

BACKGROUND AND AIMS

The complete pathological response (pCR) following neoadjuvant chemotherapy (NAC) in breast cancer is essential for the accurate prediction of prognosis. We aimed to evaluate the efficacy of the presence and type of peritumoral edema detected by magnetic resonance imaging (MRI) in predicting pCR to NAC in breast cancer patients.

METHODS AND RESULTS

One hundred five patients with the diagnosis of invasive carcinoma were evaluated by MRI before NAC. Edema was evaluated in fat-suppressed T2-weighted images. The patients were categorized into three groups: patients with no peritumoral edema, patients with peritumoral edema, and patients demonstrating subcutaneous edema. The cases were categorized as being pCR and non-pCR. Molecular subtypes, lymphovascular invasion (LVI), tumor size, and apparent diffusion coefficient (ADC) were evaluated. A positive relationship was found between the presence of edema and tumor size. Subcutaneous edema was found to be statistically higher in non-pCR patients. While the number of pCR patients with subcutaneous edema was 17 (30.4%), the number of non-pCR patients with subcutaneous edema was 26 (53.1%) (p  =  0.018). LVI was found to be statistically higher in patients with edema. The number of edema-negative and LVI (+) patients was 4 (15.4%), while the number of edema-positive and LVI (+) patients was 28 (35.4%) (p = 0.042). Intratumoral and peritumoral ADC values were significantly higher in tumors with edema.

CONCLUSION

The presence of subcutaneous edema and LVI may be utilized for the prediction of pCR outcomes in breast cancer patients scheduled for NAC treatment.

Radiomics Based on Dynamic Contrast-Enhanced MRI to Early Predict Pathologic Complete Response in Breast Cancer Patients Treated with Neoadjuvant Therapy

RATIONALE AND OBJECTIVES

To investigate the value of dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI)-based radiomics at baseline and after two cycles of neoadjuvant therapy (NAT) and associated longitudinal changes for early prediction of the NAT response in patients with breast cancer.

MATERIALS AND METHODS

One hundred seventeen patients with breast cancer who underwent DCE-MRI before NAT and after two cycles of NAT from April 2019 to November 2021 were enrolled retrospectively. Patients were randomly divided into a training set (n = 81) and a test set (n = 36) at a ratio of 7:3. Clinical-pathological data and the relative tumor maximum diameter regression value (diameter%) were also collected. A total of 851 radiomic features were extracted from the phase with the most pronounced tumor enhancement on DCE-MRI T1 imaging acquired both pre- and post-treatment. Delta and delta% radiomics features were also calculated. The Least Absolute Shrinkage and Selection Operator (LASSO) method was applied to select features, and a logistic regression model was used to calculate pre-NAT, early-NAT, delta, and delta% radscores and then select among four radscores to build a Fusion radiomics model. The final clinical-radiomics model was constructed by combining fusion radscores and clinical-pathological variables. The discrimination and clinical utility of the models were further evaluated and compared.

RESULTS

The area under the curve (AUC) values of the fusion radiomics model based on pre-NAT, Delta, and Delta% radscores were 0.868 of 0.825. The clinical-radiomics model integrating Fusion radscores and clinical-pathological variables achieved AUC values of 0.920 of 0.884, which were higher than those of the clinical model constructed by AUC values (0.858/0.831), although no significant improvement was observed in the test set (Delong test, p = 0.196). Decision curve analysis (DCA) showed that the clinical-radiomics model demonstrated more clinical utility than the clinical model.

CONCLUSION

DCE-MRI-based radiomics features may have potential for pathological complete response (pCR) prediction in the early phase of NAT. By combining radiomics features and clinical-pathological characteristics, higher diagnostic performance can be achieved.

Predicting pathological response of locally advanced breast cancer to neoadjuvant chemotherapy: comparing the performance of whole body 18F-FDG PETCT versus DCE-MRI of the breast

Background

With the expansion of the use of the neoadjuvant chemotherapy(NAC) in locally advanced breast cancer (LABC), both dynamic contrast enhanced magnetic resonance imaging (DCE-MRI) and 18F-fluorodeoxyglucose positron emission tomography (18F-FDG PET CT) are promising methods for assessment of the tumor response during chemotherapy. We aimed to evaluate the diagnostic accuracy of DCE-MRI of breast &18 F-FDG PETCT regarding the assessment of early response to neoadjuvant chemotherapy (NAC) in locally advanced breast cancer patients (LABC) and pathologic complete response (pCR) prediction.

Results

A total of forty LABC patients who had NAC were included in the study. Before and during NAC, PET/CT and DCE-MRI were used. Various morphological and functional criteria were compared and linked with post-operative pathology for both. The MRI sensitivity and specificity in assessing NAC response in conjunction with pathological data were 100% (p = 0.001) and 12.5% (p = 0.18) respectively. The equivalent readings for PET/CT were 94.1% (p = 0.001) and 25% (p = 0.18), respectively, although the estimated total accuracy for both MRI and PETCT was the same measuring 94.1% (p = 0.001) and 25% (p = 0.18) (72%). PETCT had a higher overall accuracy than MRI in assessing the response of axillary lymph nodes (ALN) to NAC (64% and 56%, respectively). Longest diameter of lesion, ADC value, and maximal enhancement in baseline MRI, SUVmax and SUV mean in baseline PETCT were all significant predictors of rCR.

Conclusion

During NAC in the primary breast mass and ALN, DCE-MRI demonstrated a better sensitivity in predicting pCR in LABC patients. Although both MRI and PETCT were equally accurate in detecting pCR of LABC patients to NAC, PETCT was more accurate in detecting pathological response of ALN to NAC.

Predicting the Early Response to Neoadjuvant Therapy with Breast MR Morphological, Functional and Relaxometry Features-A Pilot Study

Aim: To evaluate the role of MR relaxometry and derived proton density analysis in the prediction of early treatment response after two cycles of neoadjuvant therapy (NAT), in patients with breast cancer. Methods: This was a prospective study that included 59 patients with breast cancer, who underwent breast MRI prior (MRI1) and after two cycles of NAT (MRI2). The MRI1 included a sequential acquisition with five different TE’s (50, 100, 150, 200 and 250 ms) and a TR of 5000 ms. Post-processing was used to obtain the T2 relaxometry map from the MR acquisition. The tumor was delineated and seven relaxometry and proton density parameters were extracted. Additional histopathology data, T2 features and ADC were included. The response to NAT was reported based on the MRI2 as responders: partial response (>30% decreased size) and complete response (no visible tumor stable disease (SD); and non-responders: stable disease or progression (>20% increased size). Statistics was done using Medcalc software. Results: There were 50 (79.3%) patients with response and 13 (20.7%) non-responders to NAT. Age, histologic type, “in situ” component, tumor grade, estrogen and progesterone receptors, ki67% proliferation index and HER2 status were not associated with NAT response (all p > 0.05). The nodal status (N) 0 was associated with early response, while N2 was associated with non-response (p = 0.005). The tumor (T) and metastatic (M) stage were not statistically significant associated with response (p > 0.05). The margins, size and ADC values were not associated with NAT response (p-value > 0.05). The T2 min relaxometry value was associated with response (p = 0.017); a cut-off value of 53.58 obtained 86% sensitivity (95% CI 73.3−94.2), 69.23 specificity (95% CI 38.6−90.9), with an AUC = 0.715 (p = 0.038). The combined model (T2 min and N stage) achieved an AUC of 0.826 [95% CI: 0.66−0.90, p-value < 0.001]. Conclusions: MR relaxometry may be a useful tool in predicting early treatment response to NAT in breast cancer patients.

Diffusion-Weighted MRI for Predicting Pathologic Complete Response in Neoadjuvant Immunotherapy

This study tested the hypothesis that a change in the apparent diffusion coefficient (ADC) measured in diffusion-weighted MRI (DWI) is an independent imaging marker, and ADC performs better than functional tumor volume (FTV) for assessing treatment response in patients with locally advanced breast cancer receiving neoadjuvant immunotherapy. A total of 249 patients were randomized to standard neoadjuvant chemotherapy with pembrolizumab (pembro) or without pembrolizumab (control). DCE-MRI and DWI, performed prior to and 3 weeks after the start of treatment, were analyzed. Percent changes of tumor ADC metrics (mean, 5th to 95th percentiles of ADC histogram) and FTV were evaluated for the prediction of pathologic complete response (pCR) using a logistic regression model. The area under the ROC curve (AUC) estimated for the percent change in mean ADC was higher in the pembro cohort (0.73, 95% confidence interval [CI]: 0.52 to 0.93) than in the control cohort (0.63, 95% CI: 0.43 to 0.83). In the control cohort, the percent change of the 95th percentile ADC achieved the highest AUC, 0.69 (95% CI: 0.52 to 0.85). In the pembro cohort, the percent change of the 25th percentile ADC achieved the highest AUC, 0.75 (95% CI: 0.55 to 0.95). AUCs estimated for percent change of FTV were 0.61 (95% CI: 0.39 to 0.83) and 0.66 (95% CI: 0.47 to 0.85) for the pembro and control cohorts, respectively. Tumor ADC may perform better than FTV to predict pCR at an early treatment time-point during neoadjuvant immunotherapy.

Delta-Radiomics Based on Dynamic Contrast-Enhanced MRI Predicts Pathologic Complete Response in Breast Cancer Patients Treated with Neoadjuvant Chemotherapy

Simple Summary

Neoadjuvant chemotherapy (NAC) followed with surgery is the standard strategy in the treatment of locally advanced breast cancer, but the individual efficacy varies. Early and accurate prediction of complete responders determines the NAC regimens and prognosis. Breast MRI has been recommended to monitor NAC response before, during, and after treatment. Radiomics has been heralded as a breakthrough in medicine and regarded to have changed the landscape of biomedical research in oncology. Delta-radiomics characterizing the change in feature values by applying radiomics to multiple time points, is a promising strategy for predicting response after NAC. In our study, the delta-radiomics model built with the change of radiomic features before and after one cycle NAC could effectively predict pathological complete response (pCR) in breast cancer. The model provides strong support for clinical decision-making at the earliest stage and helps patients benefit the most from NAC.

Abstract

Objective: To investigate the value of delta-radiomics after the first cycle of neoadjuvant chemotherapy (NAC) using dynamic contrast-enhanced (DCE) MRI for early prediction of pathological complete response (pCR) in patients with breast cancer. Methods: From September 2018 to May 2021, a total of 140 consecutive patients (training, n = 98: validation, n = 42), newly diagnosed with breast cancer who received NAC before surgery, were prospectively enrolled. All patients underwent DCE-MRI at pre-NAC (pre-) and after the first cycle (1st-) of NAC. Radiomic features were extracted from the postcontrast early, peak, and delay phases. Delta-radiomics features were computed in each contrast phases. Least absolute shrinkage and selection operator (LASSO) and a logistic regression model were used to select features and build models. The model performance was assessed by receiver operating characteristic (ROC) analysis and compared by DeLong test. Results: The delta-radiomics model based on the early phases of DCE-MRI showed a highest AUC (0.917/0.842 for training/validation cohort) compared with that using the peak and delay phases images. The delta-radiomics model outperformed the pre-radiomics model (AUC = 0.759/0.617, p = 0.011/0.047 for training/validation cohort) in early phase. Based on the optimal model, longitudinal fusion radiomic models achieved an AUC of 0.871/0.869 in training/validation cohort. Clinical-radiomics model generated good calibration and discrimination capacity with AUC 0.934 (95%CI: 0.882, 0.986)/0.864 (95%CI: 0.746, 0.982) for training and validation cohort. Delta-radiomics based on early contrast phases of DCE-MRI combined clinicopathology information could predict pCR after one cycle of NAC in patients with breast cancer.

Robustness Evaluation of a Deep Learning Model on Sagittal and Axial Breast DCE-MRIs to Predict Pathological Complete Response to Neoadjuvant Chemotherapy

To date, some artificial intelligence (AI) methods have exploited Dynamic Contrast-Enhanced Magnetic Resonance Imaging (DCE-MRI) to identify finer tumor properties as potential earlier indicators of pathological Complete Response (pCR) in breast cancer patients undergoing neoadjuvant chemotherapy (NAC). However, they work either for sagittal or axial MRI protocols. More flexible AI tools, to be used easily in clinical practice across various institutions in accordance with its own imaging acquisition protocol, are required. Here, we addressed this topic by developing an AI method based on deep learning in giving an early prediction of pCR at various DCE-MRI protocols (axial and sagittal). Sagittal DCE-MRIs refer to 151 patients (42 pCR; 109 non-pCR) from the public I-SPY1 TRIAL database (DB); axial DCE-MRIs are related to 74 patients (22 pCR; 52 non-pCR) from a private DB provided by Istituto Tumori “Giovanni Paolo II” in Bari (Italy). By merging the features extracted from baseline MRIs with some pre-treatment clinical variables, accuracies of 84.4% and 77.3% and AUC values of 80.3% and 78.0% were achieved on the independent tests related to the public DB and the private DB, respectively. Overall, the presented method has shown to be robust regardless of the specific MRI protocol.

Advances in Imaging in Evaluating the Efficacy of Neoadjuvant Chemotherapy for Breast Cancer

Neoadjuvant chemotherapy (NAC) is increasingly widely used in breast cancer treatment, and accurate evaluation of its response provides essential information for treatment and prognosis. Thus, the imaging tools used to quantify the disease response are critical in evaluating and managing patients treated with NAC. We discussed the recent progress, advantages, and disadvantages of common imaging methods in assessing the efficacy of NAC for breast cancer.

Contrast-free MRI quantitative parameters for early prediction of pathological response to neoadjuvant chemotherapy in breast cancer

OBJECTIVES

To assess early changes in synthetic relaxometry after neoadjuvant chemotherapy (NAC) for breast cancer and establish a model with contrast-free quantitative parameters for early prediction of pathological response.

METHODS

From March 2019 to January 2021, breast MRI were performed for a primary cohort of women with breast cancer before (n = 102) and after the first (n = 93) and second (n = 90) cycle of NAC. Tumor size, synthetic relaxometry (T1/T2 relaxation time [T1/T2], proton density), and ADC were obtained, and the changes after treatment were calculated. Prediction models were established by multivariate logistic regression; evaluated with discrimination, calibration, and clinical application; and compared with Delong tests, net reclassification (NRI), and integrated discrimination index (IDI). External validation was performed from February to June 2021 with an independent cohort of 35 patients.

RESULTS

In the primary cohort, all parameters changed after early treatment. Synthetic relaxometry decreased to a greater degree in major histologic responders (MHR, Miller-Payne G4-5) compared with non-MHR (Miller-Payne G1-3). A model combining ADC after treatment, changes in T1 and tumor size, and cancer subtype achieved the highest AUC after the first (primary/validation cohort, 0.83/0.82) and second cycles (primary/validation cohort, 0.85/0.84). No difference of AUC (p ≥ 0.27), NRI (p ≥ 0.31), and IDI (p ≥ 0.32) was found between models with different cycles and size-measured sequences. Model calibration and decision curves demonstrated a good fitness and clinical benefit, respectively.

CONCLUSIONS

Early reduction in synthetic relaxometry indicated pathological response to NAC. Contrast-free T1 and ADC combined with size and cancer subtype predicted effectively pathological response after one NAC cycle.

KEY POINTS

• Synthetic MRI relaxometry changed after early neoadjuvant chemotherapy, which demonstrated pathological response for mass-like breast cancers. • Contrast-free quantitative parameters including T1 relaxation time and apparent diffusion coefficient, combined with tumor size and cancer subtype, stratified major histologic responders. • A contrast-free model predicted an early pathological response after the first treatment cycle of neoadjuvant chemotherapy.

Using the apparent diffusion coefficient histogram analysis to predict response to neoadjuvant chemotherapy in patients with breast cancer: comparison among three region of interest selection methods

Background

The apparent diffusion coefficient (ADC) value using histogram analysis is helpful to predict responses to neoadjuvant chemotherapy (NAC) in breast cancer. However, the measurement method has not reached a consensus. This study was to assess the diagnostic performance of the ADC histogram analysis at predicting patient response prior to NAC in breast cancer patients using different region of interest (ROI) selection methods.

Methods

A total of 75 patients who underwent diffusion weighted imaging (DWI) prior to NAC were retrospectively enrolled from February 2017 to December 2019. Images were measured using small 2-dimensional (2D) ROI, large 2D ROI, and volume ROI methods. The measurement time and ROI size were recorded. Histopathologic responses were acquired using the Miller-Payne grading system after surgery. The inter- and intra-observer repeatability was analyzed and the ADC histogram values from the three ROI methods were compared. The efficacy of each method at predicting patient response prior to NAC was assessed using the area under the receiver operating characteristic curve (AUC) for the whole study population and subgroups according to molecular subtype.

Results

Among the 75 enrolled patients, 26 (34.67%) were responsive to NAC therapy. The ADC histogram values were significantly different among the three ROI methods (P≤0.038). Inter- and intra-observer repeatability of the large 2D ROI method and the volume ROI method was generally greater than that observed with the 2D ROI method. The 10% ADC value of the large 2D ROI method showed the greatest AUC (0.701) in the whole study population and in the luminal subgroup (AUC 0.804). The volume ROI method required significantly more time than the other two ROI methods (P<0.001).

Conclusions

The small 2D ROI method is not appropriate for predicting response prior to NAC in breast cancer patients due to the poor repeatability. When choosing the ROI method and the histogram parameters for predicting response prior to NAC in breast cancer patients using ADC-derived histogram analysis, 10% of the large 2D ROI method is recommended, especially in luminal A subtype patients.

Value of diffusion-weighted MRI in predicting early response to neoadjuvant chemotherapy of breast cancer: comparison between ROI-ADC and whole-lesion-ADC measurements

OBJECTIVE

The aim of the study was to assess DWI with ROI-ADC and WL-ADC measurements in early response after NAC in breast cancer.

METHODS

Between January 2016 and December 2019, 55 women were enrolled in this prospective single-center study. MRI was performed at three time points for each patient: before treatment (MRI 1: DW and DCE MRI), after one cycle of NAC (MRI 2: noncontrast DW MRI), and after completion of NAC before surgery (MRI 3: DW and DCE MRI). ROI-ADC and WL-ADC measurements were obtained on MRI and were compared to histology findings and to the RCB class. Patients were categorized as having pCR or non-pCR.

RESULTS

Among 48 patients, 9 experienced pCR. An increase of ROI-ADC between MRI 1 and 2 of more than 47.5% had a sensitivity of 88.9% and a specificity of 63.4% in predicting pCR, whereas WL-ADC did not predict pCR. An increase of ROI-ADC between MRI 1 and 2 of more than 47.5% had a sensitivity of 83.3% and a specificity of 64.9% in predicting radiologic complete response. An increase of WL-ADC between MRI 1 and 2 of more than 25.5% had a sensitivity of 83.3% and a specificity of 75.5% in predicting radiologic complete response.

CONCLUSION

After one cycle of NAC, a significant increase in breast tumor ROI-ADC at DWI predicted complete pathologic and radiologic responses.

KEY POINTS

• An increase of WL-ADC between MRI 1 and 2 of more than 25.5% had a sensitivity of 83.3% and a specificity of 75.5% in predicting radiologic complete response. • An increase of ROI-ADC between MRI 1 and 2 of more than 47.5% had a sensitivity of 88.9% and a specificity of 63.4% in predicting pCR, and a sensitivity of 83.3% and a specificity of 64.9% in predicting radiologic complete response. • A significant increase in breast tumor ROI-ADC at DWI predicted complete pathologic and radiologic responses.

Nomogram for Early Prediction of Pathological Complete Response to Neoadjuvant Chemotherapy in Breast Cancer Using Dynamic Contrast-enhanced and Diffusion-weighted MRI

RATIONALE AND OBJECTIVES

The study investigated the potential of the combined use of dynamic contrast-enhanced MRI and diffusion-weighted imaging in predicting the pathological complete response (pCR) of neoadjuvant chemotherapy (NAC) after two cycles of NAC.

MATERIALS AND METHODS

Eighty-seven patients with breast cancer who underwent MR examination before and after two cycles of NAC were enrolled. The patients were randomly assigned to a training cohort and a validation cohort (3:1 ratio). MRI parameters including tumor longest diameter, time-signal intensity curve, early enhanced ratio (E₉₀), maximal enhanced ratio and ADC value were measured, and percentage change in MRI parameters were calculated. Univariate analysis and multivariate logistic regression analysis were used to evaluate independent predictors of pCR in the training cohort. The validation cohort was used to test the prediction model, and the nomogram was created based on the prediction model.

RESULTS

This study demonstrated that the ADC value after two cycles of NAC (OR = 1.041, 95% CI (1.002, 1.081); p = 0.037), percentage decrease in E₉₀ (OR = 0.927, 95% CI (0.881, 0.977); p =0.004) and percentage decrease in tumor size (OR = 0.948, 95% CI (0.909, 0.988); p = 0.011) were significantly important for independently predicting pCR. The prediction model yielded AUC of 0.939 and 0.944 in the training cohort and the validation cohort, respectively.

CONCLUSION

The combined use of dynamic contrast-enhanced MRI and diffusion-weighted imaging could accurately predict pCR after two cycles of NAC. The prediction model and the nomogram had strong predictive value to NAC.

Factors affecting the value of diffusion-weighted imaging for identifying breast cancer patients with pathological complete response on neoadjuvant systemic therapy: a systematic review

This review aims to identify factors causing heterogeneity in breast DWI-MRI and their impact on its value for identifying breast cancer patients with pathological complete response (pCR) on neoadjuvant systemic therapy (NST). A search was performed on PubMed until April 2020 for studies analyzing DWI for identifying breast cancer patients with pCR on NST. Technical and clinical study aspects were extracted and assessed for variability. Twenty studies representing 1455 patients/lesions were included. The studies differed with respect to study population, treatment type, DWI acquisition technique, post-processing (e.g., mono-exponential/intravoxel incoherent motion/stretched exponential modeling), and timing of follow-up studies. For the acquisition and generation of ADC-maps, various b-value combinations were used. Approaches for drawing regions of interest on longitudinal MRIs were highly variable. Biological variability due to various molecular subtypes was usually not taken into account. Moreover, definitions of pCR varied. The individual areas under the curve for the studies range from 0.50 to 0.92. However, overlapping ranges of mean/median ADC-values at pre- and/or during and/or post-NST were found for the pCR and non-pCR groups between studies. The technical, clinical, and epidemiological heterogeneity may be causal for the observed variability in the ability of DWI to predict pCR accurately. This makes implementation of DWI for pCR prediction and evaluation based on one absolute ADC threshold for all breast cancer types undesirable. Multidisciplinary consensus and appropriate clinical study design, taking biological and therapeutic variation into account, is required for obtaining standardized, reliable, and reproducible DWI measurements for pCR/non-pCR identification.

Breast imaging: Beyond the detection

Breast cancer is a heterogeneous disease nowadays, including different biological subtypes with a variety of possible treatments, which aim to achieve the best outcome in terms of response to therapy and overall survival. In recent years breast imaging has evolved considerably, and the ultimate goal is to predict these strong phenotypic differences noninvasively. Indeed, breast cancer multiparametric studies can highlight not only qualitative imaging parameters, as the presence/absence of a likely malignant finding, but also quantitative parameters, suggesting clinical-pathological features through the evaluation of imaging biomarkers. A further step has been the introduction of artificial intelligence and in particular radiogenomics, that investigates the relationship between breast cancer imaging characteristics and tumor molecular, genomic and proliferation features. In this review, we discuss the main techniques currently in use for breast imaging, their respective fields of use and their technological and diagnostic innovations.

Radiomics Analysis of Multi-Phase DCE-MRI in Predicting Tumor Response to Neoadjuvant Therapy in Breast Cancer

Objective: To explore whether the pretreatment dynamic contrast enhancement magnetic resonance imaging (DCE-MRI) and radiomics signatures were associated with pathologic complete response (pCR) to neoadjuvant therapy (NAT) in breast cancer. Method: A retrospective review of 70 patients with breast invasive carcinomas proved by biopsy between June 2017 and October 2020 (26 patients were pathological complete response, and 44 patients were non-pathological complete response). Within the pre-contrast and five post-contrast dynamic series, a total of 1037 quantitative imaging features were extracted from in each phase. Additionally, the Δfeatures (the difference between the features before and after the comparison) were used for subsequent analysis. The least absolute shrinkage and selection operator (LASSO) regression method was used to select features related to pCR, and then use these features to train multiple machine learning classifiers to predict the probability of pCR for a given patient. The area under the curve (AUC), accuracy, sensitivity, and specificity were calculated to assess the predictive performances of the radiomics model for each of the five phases of time points. Result: Among the five phases, each individual phase performed with AUCs ranging from 0.845 to 0.919 in predicting pCR. The best single phases performance was given by the 3rd phase (AUC = 0.919, sensitivity 0.885, specificity 0.864). 5 of the features have significant differences between pCR and non-pCR groups in each phase, most features reach their maximum or minimum in the 2nd or 3rd phase. Conclusion: The radiomic features extracted from each phase of pre-treatment DCE-MRI possess discriminatory power to predict tumor response.

Quantitative Assessment of the Echogenicity of a Breast Tumor Predicts the Response to Neoadjuvant Chemotherapy

The aim of the study was to improve monitoring the treatment response in breast cancer patients undergoing neoadjuvant chemotherapy (NAC). The IRB approved this prospective study. Ultrasound examinations were performed prior to treatment and 7 days after four consecutive NAC cycles. Residual malignant cell (RMC) measurement at surgery was the standard of reference. Alteration in B-mode ultrasound (tumor echogenicity and volume) and the Kullback-Leibler divergence (kld), as a quantitative measure of amplitude difference, were used. Correlations of these parameters with RMC were assessed and Receiver Operating Characteristic curve (ROC) analysis was performed. Thirty-nine patients (mean age 57 y.) with 50 tumors were included. There was a significant correlation between RMC and changes in quantitative parameters (KLD) after the second, third and fourth course of NAC, and alteration in echogenicity after the third and fourth course. Multivariate analysis of the echogenicity and KLD after the third NAC course revealed a sensitivity of 91%, specificity of 92%, PPV = 77%, NPV = 97%, accuracy = 91%, and AUC of 0.92 for non-responding tumors (RMC ≥ 70%). In conclusion, monitoring the echogenicity and KLD parameters made it possible to accurately predict the treatment response from the second course of NAC.

Value of Machine Learning with Multiphases CE-MRI Radiomics for Early Prediction of Pathological Complete Response to Neoadjuvant Therapy in HER2-Positive Invasive Breast Cancer

Background

To assess the value of radiomics based on multiphases contrast-enhanced magnetic resonance imaging (CE-MRI) for early prediction of pathological complete response (pCR) to neoadjuvant therapy (NAT) in patients with human epithelial growth factor receptor 2 (HER2) positive invasive breast cancer.

Methods

A total of 127 patients with newly diagnosed primary HER2 positive invasive breast cancer underwent CE-MRI before NAT and performed surgery after NAT. Radiomic features were extracted from the 1^st postcontrast CE-MRI phase (CE₁) and multi-phases CE-MRI (CE_m),respectively. With selected features using a forward stepwise regression, 23 machine learning classifiers based on CE₁ and CE_m were constructed respectively for differentiating pCR and non-pCR patients. The performances of classifiers were assessed and compared by their accuracy, sensitivity, specificity and AUC (area under curve). The optimal machine learning classification was used to discriminate pCR vs non-pCR in mass and non-mass groups, uni-focal and unilateral multi-focal groups, respectively.

Results

For the task of pCR classification, 6 radiomic features from CE₁ and 6 from CE_m were selected for the construction of machine learning models, respectively. The linear SVM based on CE_m outperformed the logistic regression model using CE₁ with an AUC of 0.84 versus 0.69. In mass and non-mass enhancement groups, the accuracy of linear SVM achieved 84% and 76%. Whereas in unifocal and unilateral multifocal cases, 79% and 75% accuracy were achieved by linear SVM.

Conclusion

Multiphases CE-MRI imaging may offer more heterogeneity information in the tumor and provide a non-invasive approach for early prediction of pCR to NAT in patients with HER2-positive invasive breast cancer.

Early response by MR imaging and ultrasound as predictor of pathologic complete response to 12-week neoadjuvant therapy for different early breast cancer subtypes: Combined analysis from the WSG ADAPT subtrials

We evaluated the role of early response after 3 weeks of neoadjuvant treatment (NAT) assessed by ultrasound (US), magnetic resonance imaging (MRI) and Ki-67 dynamics for prediction of pathologic complete response (pCR) in different early breast cancer subtypes. Patients with HR+/HER2+, HR-/HER2- and HR-/HER2+ tumors enrolled into three neoadjuvant WSG ADAPT subtrials underwent US, MRI and Ki-67 assessment at diagnosis and after 3 weeks of NAT. Early response was defined as complete or partial response (US, MRI) and ≥30% proliferation decrease or <500 invasive tumor cells (Ki-67). Predictive values and area under the receiver operating characteristic (AUC) curves for prediction of pCR (ypT0/is ypN0) after 12-week NAT were calculated. Two hundred twenty-six had MRI and 401 US; 107 underwent both MRI and US. All three methods yielded a similar AUC in HR+/HER2+ (0.66-0.67) and HR-/HER2- tumors (0.53-0.63), while MRI and Ki-67 performed better than US in HR-/HER2+ tumors (0.83 and 0.79 vs 0.56). Adding MRI+/-Ki-67 increased AUC of US in HR-/HER2+ tumors to 0.64 to 0.75. MRI and Ki-67 demonstrated highest sensitivity in HR-/HER2- (0.8-1) and HR-/HER2+ tumors (1, both). Negative predictive value was similar for all methods in HR+/HER2+ (0.71-0.74) and HR-/HER2- tumors (0.85-1), while it was higher for MRI and Ki-67 compared to US in HR-/HER2+ subtype (1 vs 0.5). Early response assessed by US, MRI and Ki-67 is a strong predictor for pCR after 12-week NAT. Strength of pCR prediction varies according to tumor subtype. Adding MRI+/-Ki-67 to US did not improve pCR prediction in majority of our patients.

Radiomics of Tumor Heterogeneity in Longitudinal Dynamic Contrast-Enhanced Magnetic Resonance Imaging for Predicting Response to Neoadjuvant Chemotherapy in Breast Cancer

Breast tumor morphological and vascular characteristics can be changed during neoadjuvant chemotherapy (NACT). The early changes in tumor heterogeneity can be quantitatively modeled by longitudinal dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI), which is useful in predicting responses to NACT in breast cancer. In this retrospective analysis, 114 female patients with unilateral unifocal primary breast cancer who received NACT were included in a development (n = 61) dataset and a testing dataset (n = 53). DCE-MRI was performed for each patient before and after treatment (two cycles of NACT) to generate baseline and early follow-up images, respectively. Feature-level changes (delta) of the entire tumor were evaluated by calculating the relative net feature change (deltaRAD) between baseline and follow-up images. The voxel-level change inside the tumor was evaluated, which yielded a Jacobian map by registering the follow-up image to the baseline image. Clinical information and the radiomic features were fused to enhance the predictive performance. The area under the curve (AUC) values were assessed to evaluate the prediction performance. Predictive models using radiomics based on pre- and post-treatment images, Jacobian maps and deltaRAD showed AUC values of 0.568, 0.767, 0.630 and 0.726, respectively. When features from these images were fused, the predictive model generated an AUC value of 0.771. After adding the molecular subtype information in the fused model, the performance was increased to an AUC of 0.809 (sensitivity of 0.826 and specificity of 0.800), which is significantly higher than that of the baseline imaging- and Jacobian map-based predictive models (p = 0.028 and 0.019, respectively). The level of tumor heterogeneity reduction (evaluated by texture feature) is higher in the NACT responders than in the nonresponders. The results suggested that changes in DCE-MRI features that reflect a reduction in tumor heterogeneity following NACT could provide early prediction of breast tumor response. The prediction was improved when the molecular subtype information was combined into the model.

Tissue Sodium Concentration Quantification at 7.0-T MRI as an Early Marker for Chemotherapy Response in Breast Cancer: A Feasibility Study

Background Tissue sodium concentration (TSC) is elevated in breast cancer and can determine chemotherapy response. Purpose To test the feasibility of using a sodium 23 (²³Na) MRI protocol at 7.0 T for TSC quantification to predict early treatment outcomes of neoadjuvant chemotherapy in breast cancer and to determine whether those quantitative values provide additional information about efficacy. Materials and Methods Women with primary breast cancer were included in this prospective study. From July 2017 to June 2018, participants underwent 7.0-T ²³Na MRI. Multichannel data sets were acquired with a density-adapted, three-dimensional radial projection reconstruction pulse sequence. Two-dimensional tumor size and TSC were evaluated before and after the first and second chemotherapy cycle, and statistical tests were performed based on the presence or absence of a pathologic complete response (pCR). Results Fifteen women with breast cancer and six healthy women were enrolled. The mean baseline tumor size in women with a pCR was 7.0 cm² ± 5.0 (standard deviation), and the mean baseline tumor size in women without a pCR was 19.0 cm² ± 12.0. After the first chemotherapy cycle, women with a pCR showed a reduced tumor size of 32.9% (2.3 cm²/7.0 cm²), compared with 15.3% (2.9 cm²/19.0 cm²) in those without a pCR. The areas under the receiver operating characteristic curve for tumor size reduction after the first and second chemotherapy cycle were 0.73 (95% CI: 0.09, 0.50; P = .12) and 0.93 (95% CI: 0.04, 0.60; P < .001), respectively. Women with a pCR had a mean baseline TSC of 69.4 mmol/L ± 6.1, with a reduction of 12.0% (8.3 mmol/L), whereas those without a pCR had a mean baseline TSC of 71.7 mmol/L ± 5.7, with a reduction of 4.7% (3.4 mmol/L) after the first cycle. The areas under the receiver operating characteristic curve for TSC after the first and second cycles were 0.96 (95% CI: 0.86, 1.00; P < .001) and 1.000 (95% CI: 1.00, P < .001), respectively. Conclusion Using 7.0-T MRI for tissue sodium concentration quantification to predict early treatment outcomes of neoadjuvant chemotherapy in breast cancer is feasible, with reduced tissue sodium concentration indicative of cancer response. © RSNA, 2021 Online supplemental material is available for this article.

Diffusion MRI of the breast: Current status and future directions

Diffusion-weighted imaging (DWI) is increasingly being incorporated into routine breast MRI protocols in many institutions worldwide, and there are abundant breast DWI indications ranging from lesion detection and distinguishing malignant from benign tumors to assessing prognostic biomarkers of breast cancer and predicting treatment response. DWI has the potential to serve as a noncontrast MR screening method. Beyond apparent diffusion coefficient (ADC) mapping, which is a commonly used quantitative DWI measure, advanced DWI models such as intravoxel incoherent motion (IVIM), non-Gaussian diffusion MRI, and diffusion tensor imaging (DTI) are extensively exploited in this field, allowing the characterization of tissue perfusion and architecture and improving diagnostic accuracy without the use of contrast agents. This review will give a summary of the clinical literature along with future directions. Level of Evidence: 5 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2020;52:70-90.

Pretreatment apparent diffusion coefficient does not predict therapy response to neoadjuvant chemotherapy in breast cancer

Background

Some reports indicated that apparent diffusion coefficient can predict pathologic response to treatment in breast cancer (BC). The purpose of the present meta-analysis was to provide evident data regarding use of ADC values for prediction of treatment response in BC.

Methods

MEDLINE library, EMBASE and SCOPUS databases were screened for associations between ADC and treatment response for neoadjuvant chemotherapy in breast cancer (BC) up to March 2020. Overall, 22 studies met the inclusion criteria. For the present analysis, the following data were extracted from the collected studies: authors, year of publication, study design, number of patients/lesions, mean and standard deviation of the pretreatment ADC values. The methodological quality of the included studies was checked according to the QUADAS-2 instrument. The meta-analysis was undertaken by using RevMan 5.3 software. DerSimonian and Laird random-effects models with inverse-variance weights were used without any further correction to account for the heterogeneity between the studies. Mean ADC values including 95% confidence intervals were calculated separately for responders and non responders.

Results

The acquired 22 studies comprised 1827 patients with different BC. Of the 1827 patients, 650 (35.6%) were reported as responders and 1177 (64.4%) as non-responders to the neoadjuvant chemotherapy. The pooled calculated pretreatment mean ADC value of BC in responders was 0.98 (95% CI = [0.94; 1.03]). In non-responders, it was 1.05 (95% CI = [1.00; 1.10]). The ADC values of the groups overlapped significantly.

Conclusion

Pretreatment ADC alone cannot predict response to neoadjuvant chemotherapy in BC.

The additive role of 1H-magnetic resonance spectroscopic imaging to ensure pathological complete response after neoadjuvant chemotherapy in breast cancer patients

Purpose

To assess the role of ¹H-magnetic resonance spectroscopy (¹H-MRS) in the confirmation of pathological complete response after neoadjuvant chemotherapy in breast cancer.

Material and methods

Forty-seven cases (53.72 ± 8.53 years) were evaluated using magnetic resonance imaging (MRI) and ¹H-MRS with choline (Cho) signal-to-noise ratio (SNR) measured followed by histopathology and ROC analyses.

Results

Twelve patients had complete response, and 35 patients had residual disease. Mean age was 53.72 ± 8.53 years. The mean tumour size before neoadjuvant chemotherapy (NAC) was 4.21 ± 0.99 cm and after NAC was 0.9 ± 0.44 cm.Positive total choline signal (tCho) was detected in all cases. The mean Cho SNR before NAC was 9.53 ± 1.7 and after NAC was 2.53 ± 1.3. The Cho SNR cut-off point differentiating between pathologic complete response (pCR) and the non pCR was 1.95. Dynamic MRI showed 83.3% sensitivity, 65.7% specificity, 45.5% positive predictive value, 92.0% negative predictive value, and 70.2% diagnostic accuracy. Combined evaluation done by using the dynamic MRI and ¹H-MRS showed 91.5% diagnostic accuracy with 75.0% sensitivity, 97.1% specificity, 75% positive predictive value, and 91.9% negative predictive value. ROC curves of Cho SNR showed statistically significant differences between non pCR and pCR with AUC was 0.955, 82.9% sensitivity, 91.7% specificity, 96.7% positive predictive value, 64.7% negative predictive value, and 85.11% diagnostic accuracy.

Conclusions

¹H-MRS improves the diagnostic accuracy in the prediction of the pCR after NAC.

Machine learning with multiparametric magnetic resonance imaging of the breast for early prediction of response to neoadjuvant chemotherapy

In patients with locally advanced breast cancer undergoing neoadjuvant chemotherapy (NAC), some patients achieve a complete pathologic response (pCR), some achieve a partial response, and some do not respond at all or even progress. Accurate prediction of treatment response has the potential to improve patient care by improving prognostication, enabling de-escalation of toxic treatment that has little benefit, facilitating upfront use of novel targeted therapies, and avoiding delays to surgery. Visual inspection of a patient's tumor on multiparametric MRI is insufficient to predict that patient's response to NAC. However, machine learning and deep learning approaches using a mix of qualitative and quantitative MRI features have recently been applied to predict treatment response early in the course of or even before the start of NAC. This is a novel field but the data published so far has shown promising results. We provide an overview of the machine learning and deep learning models developed to date, as well as discuss some of the challenges to clinical implementation.

Joint Prediction of Breast Cancer Histological Grade and Ki-67 Expression Level Based on DCE-MRI and DWI Radiomics

OBJECTIVE

Histologic grade and Ki-67 proliferation status are important clinical indictors for breast cancer prognosis and treatment. The purpose of this study is to improve prediction accuracy of these clinical indicators based on tumor radiomic analysis.

METHODS

We jointly predicted Ki-67 and tumor grade with a multitask learning framework by separately utilizing radiomics from tumor MRI series. Additionally, we showed how multitask learning models (MTLs) could be extended to combined radiomics from the MRI series for a better prediction based on the assumption that features from different sources of images share common patterns while providing complementary information. Tumor radiomic analysis was performed with morphological, statistical and textural features extracted on the DWI and dynamic contrast-enhanced MRI (DCE-MRI) series of the precontrast and subtraction images, respectively.

RESULTS

Joint prediction of Ki-67 status and tumor grade on MR images using the MTL achieved performance improvements over that of single-task-based predictive models. Similarly, for the prediction tasks of Ki-67 and tumor grade, the MTL for combined precontrast and apparent diffusion coefficient (ADC) images achieved AUCs of 0.811 and 0.816, which were significantly better than that of the single-task- based model with p values of 0.005 and 0.017, respectively.

CONCLUSION

Mapping MRI radiomics to two related clinical indicators improves prediction performance for both Ki-67 expression level and tumor grade.

SIGNIFICANCE

Joint prediction of indicators by multitask learning that combines correlations of MRI radiomics is important for optimal tumor therapy and treatment because clinical decisions are made by integrating multiple clinical indicators.

Impact of Machine Learning With Multiparametric Magnetic Resonance Imaging of the Breast for Early Prediction of Response to Neoadjuvant Chemotherapy and Survival Outcomes in Breast Cancer Patients

PURPOSE

The aim of this study was to assess the potential of machine learning with multiparametric magnetic resonance imaging (mpMRI) for the early prediction of pathological complete response (pCR) to neoadjuvant chemotherapy (NAC) and of survival outcomes in breast cancer patients.

MATERIALS AND METHODS

This institutional review board-approved prospective study included 38 women (median age, 46.5 years; range, 25-70 years) with breast cancer who were scheduled for NAC and underwent mpMRI of the breast at 3 T with dynamic contrast-enhanced (DCE), diffusion-weighted imaging (DWI), and T2-weighted imaging before and after 2 cycles of NAC. For each lesion, 23 features were extracted: qualitative T2-weighted and DCE-MRI features according to BI-RADS (Breast Imaging Reporting and Data System), quantitative pharmacokinetic DCE features (mean plasma flow, volume distribution, mean transit time), and DWI apparent diffusion coefficient (ADC) values. To apply machine learning to mpMRI, 8 classifiers including linear support vector machine, linear discriminant analysis, logistic regression, random forests, stochastic gradient descent, decision tree, adaptive boosting, and extreme gradient boosting (XGBoost) were used to rank the features. Histopathologic residual cancer burden (RCB) class (with RCB 0 being a pCR), recurrence-free survival (RFS), and disease-specific survival (DSS) were used as the standards of reference. Classification accuracy with area under the receiving operating characteristic curve (AUC) was assessed using all the extracted qualitative and quantitative features for pCR as defined by RCB class, RFS, and DSS using recursive feature elimination. To overcome overfitting, 4-fold cross-validation was used.

RESULTS

Machine learning with mpMRI achieved stable performance as shown by mean classification accuracies for the prediction of RCB class (AUC, 0.86) and DSS (AUC, 0.92) based on XGBoost and the prediction of RFS (AUC, 0.83) with logistic regression. The XGBoost classifier achieved the most stable performance with high accuracies compared with other classifiers. The most relevant features for the prediction of RCB class were as follows: changes in lesion size, complete pattern of shrinkage, and mean transit time on DCE-MRI; minimum ADC on DWI; and peritumoral edema on T2-weighted imaging. The most relevant features for prediction of RFS were as follows: volume distribution, mean plasma flow, and mean transit time; DCE-MRI lesion size; minimum, maximum, and mean ADC with DWI. The most relevant features for prediction of DSS were as follows: lesion size, volume distribution, and mean plasma flow on DCE-MRI, and maximum ADC with DWI.

CONCLUSIONS

Machine learning with mpMRI of the breast enables early prediction of pCR to NAC as well as survival outcomes in breast cancer patients with high accuracy and thus may provide valuable predictive information to guide treatment decisions.

Accuracy of multi-parametric breast MR imaging for predicting pathological complete response of operable breast cancer prior to neoadjuvant systemic therapy

OBJECTIVES

To evaluate whether multiparametric breast-MRI, obtained before the initiation of neoadjuvant systemic therapy (NST) for operable breast cancer, predicts which cancer will achieve a pathological complete response (pCR) after the completion of NST.

METHODS

This was an IRB-approved retrospective study on 31 consecutive patients (median age, 56 years) with operable invasive breast cancer (median size: 22 mm; triple-negative: 11/31 [35%], HER2-positive: 7/31 [23%], triple-positive: 13/31 [42%]) who underwent multiparametric DCE-MRI before the initiation of NST. The MRI protocol consisted of high-resolution dynamic contrast-enhanced MRI (DCE-MRI), T2-TSE, and DWI (b-values 0, 100, 800 s/mm²). The results of surgical pathology after the completion of NST served as a standard of reference. Patient characteristics (age and menopausal status), pathological tumor characteristics (type, stage, nuclear grade, ER/PR and HER2 receptor status, and Ki-67 staining), and MRI characteristics (size, morphology, T2 signal intensity, enhancement kinetics, and ADC values) before NST were evaluated and compared between patients achieving pCR vs. non-pCR.

RESULTS

Among 31 patients, 17 achieved pCR (55%) and 14 non-pCR (45%). No correlation was observed between patient- or tumor pathology-derived characteristics and pCR vs. non-pCR. Among MRI-derived tumor characteristics, tumor growth orientation parallel to Cooper's ligaments (p = 0.002) and wash-out rates (p = 0.019) correlated with pCR. Pre-NST ADC values were lower in patients achieving pCR (P = 0.086).

CONCLUSIONS

A tumor growth pattern parallel with Cooper's ligaments and a fast wash-out rate on pre-treatment multiparametric MRI are predictive of pCR and more closely associated with pCR than ADC values.

MRI Volume Changes of Axillary Lymph Nodes as Predictor of Pathologic Complete Responses to Neoadjuvant Chemotherapy in Breast Cancer

INTRODUCTION

Longitudinal monitoring of breast tumor volume over the course of chemotherapy is informative of pathologic response. This study aims to determine whether axillary lymph node (aLN) volume by magnetic resonance imaging (MRI) could augment the prediction accuracy of treatment response to neoadjuvant chemotherapy (NAC).

MATERIALS AND METHODS

Level-2a curated data from the I-SPY-1 TRIAL (2002-2006) were used. Patients had stage 2 or 3 breast cancer. MRI was acquired pre-, during, and post-NAC. A subset with visible aLNs on MRI was identified (N = 132). Prediction of pathologic complete response (PCR) was made using breast tumor volume changes, nodal volume changes, and combined breast tumor and nodal volume changes with sub-stratification with and without large lymph nodes (3 mL or ∼1.79 cm diameter cutoff). Receiver operating characteristic curve analysis was used to quantify prediction performance.

RESULTS

The rate of change of aLN and breast tumor volume were informative of pathologic response, with prediction being most informative early in treatment (area under the curve (AUC), 0.57-0.87) compared with later in treatment (AUC, 0.50-0.75). Larger aLN volume was associated with hormone receptor negativity, with the largest nodal volume for triple negative subtypes. Sub-stratification by node size improved predictive performance, with the best predictive model for large nodes having AUC of 0.87.

CONCLUSION

aLN MRI offers clinically relevant information and has the potential to predict treatment response to NAC in patients with breast cancer.

Multiparametric MR imaging to assess response following neoadjuvant systemic treatment in various breast cancer subtypes: Comparison between different definitions of pathologic complete response

OBJECTIVES

To validate the performance of multiparametric magnetic resonance (MR) imaging to assess pathologic response to neoadjuvant systemic therapy (NST) in various breast cancer subtypes considering two definitions of pCR: absence of any residual invasive cancer or DCIS (ypT0) and absence of invasive tumour cells (ypT0/is).

METHODS

Institutional review board-approved retrospective study, with waiver of the need to obtain informed consent. From January 2015 to June 2017, 81 women with 82 breast cancers undergoing NST were included. Eighteen lesions (22%) were immunohistochemically HER2-positive, 12 (15%) triple negative (TN), 42 (51%) luminal B-like and 10 (12%) luminal B-like/HER2-positive. Breast MR imaging was performed before and after NST. A comparative analysis considering pCR as ypT0 and ypT0/is was carried out. Performance of univariate and multivariate models to potentially predict pathologic response were evaluated.

RESULTS

ypT0 was attained in 23% (19/82) of cases and ypT0/is in 33% (27/82) of cases. In both scenarios, HER2-positive subtype achieved the best response, 53% and 48%, respectively. A significant relationship was found between late enhancement and pathologic response (p < 0.001) regardless of pCR definition. In the ypT0 scenario, mean ADC ratio in the pCR subgroup was significantly higher than that in the non-pCR subgroup (p = 0.021) but no significant relationship was noted in ypT0/is. A multivariate model including MR late enhancement, ADC ratio and tumor subtype identified pathologic response with 86% and 84% accuracy when ypT0 and ypT0/is were considered, respectively.

CONCLUSION

MR imaging late enhancement and ADC ratio along with breast cancer IHC subtype identify pathologic response following NST with high accuracy, achieving the highest NPV in TN and HER2-positive tumors and the highest PPV in luminal B-like subtypes, regardless of the definition of pCR as ypT0 or ypT0/is. In light of these findings and given that residual DCIS does not have an impact on survival rates, ypT0/is seems to be the preferable definition of pCR.

Additive value of diffusion-weighted MRI in the I-SPY 2 TRIAL

BACKGROUND

The change in apparent diffusion coefficient (ADC) measured from diffusion-weighted imaging (DWI) has been shown to be predictive of pathologic complete response (pCR) for patients with locally invasive breast cancer undergoing neoadjuvant chemotherapy.

PURPOSE

To investigate the additive value of tumor ADC in a multicenter clinical trial setting.

STUDY TYPE

Retrospective analysis of multicenter prospective data.

POPULATION

In all, 415 patients who enrolled in the I-SPY 2 TRIAL from 2010 to 2014 were included.

FIELD STRENGTH/SEQUENCE

1.5T or 3T MRI system using a fat-suppressed single-shot echo planar imaging sequence with b-values of 0 and 800 s/mm2 for DWI, followed by a T1-weighted sequence for dynamic contrast-enhanced MRI (DCE-MRI) performed at pre-NAC (T0), after 3 weeks of NAC (T1), mid-NAC (T2), and post-NAC (T3).

ASSESSMENT

Functional tumor volume and tumor ADC were measured at each MRI exam; pCR measured at surgery was assessed as the binary outcome. Breast cancer subtype was defined by hormone receptor (HR) and human epidermal growth factor receptor 2 (HER2) status.

STATISTICAL TESTS

A logistic regression model was used to evaluate associations between MRI predictors with pCR. The cross-validated area under the curve (AUC) was calculated to assess the predictive performance of the model with and without ADC.

RESULTS

In all, 354 patients (128 HR+/HER2-, 60 HR+/HER2+, 34 HR-/HER2+, 132 HR-/HER2-) were included in the analysis. In the full cohort, adding ADC predictors increased the AUC from 0.76 to 0.78 at mid-NAC and from 0.76 to 0.81 at post-NAC. In HR/HER2 subtypes, the AUC increased from 0.52 to 0.65 at pre-NAC for HR+/HER2-, from 0.67 to 0.73 at mid-NAC and from 0.72 to 0.76 at post-NAC for HR+/HER2+, from 0.71 to 0.81 at post-NAC for triple negatives.

DATA CONCLUSION

The addition of ADC to standard functional tumor volume MRI showed improvement in the prediction of treatment response in HR+ and triple-negative breast cancer.

LEVEL OF EVIDENCE

2 Technical Efficacy Stage: 4 J. Magn. Reson. Imaging 2019;50:1742-1753.

Diffusion-weighted MRI Findings Predict Pathologic Response in Neoadjuvant Treatment of Breast Cancer: The ACRIN 6698 Multicenter Trial

Purpose To determine if the change in tumor apparent diffusion coefficient (ADC) at diffusion-weighted (DW) MRI is predictive of pathologic complete response (pCR) to neoadjuvant chemotherapy for breast cancer. Materials and Methods In this prospective multicenter study, 272 consecutive women with breast cancer were enrolled at 10 institutions (from August 2012 to January 2015) and were randomized to treatment with 12 weekly doses of paclitaxel (with or without an experimental agent), followed by 12 weeks of treatment with four cycles of anthracycline. Each woman underwent breast DW MRI before treatment, at early treatment (3 weeks), at midtreatment (12 weeks), and after treatment. Percentage change in tumor ADC from that before treatment (ΔADC) was measured at each time point. Performance for predicting pCR was assessed by using the area under the receiver operating characteristic curve (AUC) for the overall cohort and according to tumor hormone receptor (HR)/human epidermal growth factor receptor 2 (HER2) disease subtype. Results The final analysis included 242 patients with evaluable serial imaging data, with a mean age of 48 years ± 10 (standard deviation); 99 patients had HR-positive (hereafter, HR+)/HER2-negative (hereafter, HER2-) disease, 77 patients had HR-/HER2- disease, 42 patients had HR+/HER2+ disease, and 24 patients had HR-/HER2+ disease. Eighty (33%) of 242 patients experienced pCR. Overall, ΔADC was moderately predictive of pCR at midtreatment/12 weeks (AUC = 0.60; 95% confidence interval [CI]: 0.52, 0.68; P = .017) and after treatment (AUC = 0.61; 95% CI: 0.52, 0.69; P = .013). Across the four disease subtypes, midtreatment ΔADC was predictive only for HR+/HER2- tumors (AUC = 0.76; 95% CI: 0.62, 0.89; P < .001). In a test subset, a model combining tumor subtype and midtreatment ΔADC improved predictive performance (AUC = 0.72; 95% CI: 0.61, 0.83) over ΔADC alone (AUC = 0.57; 95% CI: 0.44, 0.70; P = .032.). Conclusion After 12 weeks of therapy, change in breast tumor apparent diffusion coefficient at MRI predicts complete pathologic response to neoadjuvant chemotherapy. © RSNA, 2018 Online supplemental material is available for this article.

Diffusion-weighted MR imaging of locally advanced breast carcinoma: the optimal time window of predicting the early response to neoadjuvant chemotherapy

Background

It is very difficult to predict the early response to NAC only on the basis of change in tumor size. ADC value derived from DWI promises to be a valuable parameter for evaluating the early response to treatment. This study aims to establish the optimal time window of predicting the early response to neoadjuvant chemotherapy (NAC) for different subtypes of locally advanced breast carcinoma using diffusion-weighted imaging (DWI).

Methods

We conducted an institutional review board-approved prospective clinical study of 142 patients with locally advanced breast carcinoma. All patients underwent conventional MR and DW examinations prior to treatment and after first, second, third, fourth, sixth and eighth cycle of NAC. The response to NAC was classified into a pathologic complete response (pCR) and a non-pCR group. DWI parameters were compared between two groups, and the optimal time window for predicting tumor response was established for each chemotherapy regimen.

Results

For all the genomic subtypes, there were significant differences in baseline ADC value between pCR and non-pCR group (p < 0.05). The time point prior to treatment could be considered as the ideal time point regardless of genomic subtype. In the group that started with taxanes or anthracyclines, for Luminal A or Luminal B subtype, postT1 could be used as the ideal time point during chemotherapy; for Basal-like or HER2-enriched subtype, postT2 as the ideal time point during chemotherapy. In the group that started with taxanes and anthracyclines, for HER2-enriched, Luminal B or Basal-like subtype, postT1 could be used as the ideal time point during chemotherapy; for Luminal A subtype, postT2 as the ideal time point during chemotherapy.

Conclusions

The time point prior to treatment can be considered as the optimal time point regardless of genomic subtype. For each chemotherapy regimen, the optimal time point during chemotherapy varies across different genomic subtypes.

Can Multi-Parametric MR Based Approach Improve the Predictive Value of Pathological and Clinical Therapeutic Response in Breast Cancer Patients?

The potential of total choline (tCho), apparent diffusion coefficient (ADC) and tumor volume, both individually and in combination of all these three parameters (multi-parametric approach), was evaluated in predicting both pathological and clinical responses in 42 patients with locally advanced breast cancer (LABC) enrolled for neoadjuvant chemotherapy (NACT). Patients were sequentially examined by conventional MRI; diffusion weighted imaging and in vivo proton MR spectroscopy at 4 time points (pre-therapy, after I, II, and III NACT) at 1.5 T. Miller Payne grading system was used for pathological assessment of response. Of the 42 patients, 24 were pathological responders (pR) while 18 were pathological non-responders (pNR). Clinical response determination classified 26 patients as responders (cR) while 16 as non-responders (cNR). tCho and ADC showed significant changes after I NACT, however, MR measured tumor volume showed reduction only after II NACT both in pR and cR. After III NACT, the sensitivity to detect responders was highest for MR volume (83.3% for pR and 96.2% for cR) while the specificity was highest for ADC (76.5% for pR and 100% for cR). Combination of all three parameters exhibited lower sensitivity (66.7%) than MR volume for pR prediction, however, a moderate improvement was seen in specificity (58.8%). For the prediction of clinical response, multi-parametric approach showed 84.6% sensitivity with 100% specificity compared to MR volume (sensitivity 96.2%; specificity 80%). Kappa statistics demonstrated substantial agreement of clinical response with MR volume (k = 0.78) and with multi-parametric approach (k = 0.80) while moderate agreement was seen for tCho (k = 0.48) and ADC (k = 0.46). The values of k for tCho, MR volume and ADC were 0.31, 0.38, and 0.18 indicating fair, moderate, and slight agreement, respectively with pathological response. Moderate agreement (k = 0.44) was observed between clinical and pathological responses. Our study demonstrated that both tCho and ADC are strong predictors of assessment of early pathological and clinical responses. Multi-parametric approach yielded 100% specificity in predicting clinical response. Following III NACT, MR volume emerged as highly suitable predictor for both clinical and pathological assessments. PCA demonstrated separate clusters of pR vs. pNR and cR vs. cNR at post-therapy while with some overlap at pre-therapy.

Apparent diffusion coefficient measurement in luminal breast cancer: will tumour shrinkage patterns affect its efficacy of evaluating the pathological response?

AIM

To determine which region of interest (ROI) placement method of apparent diffusion coefficient (ADC) measurement has the best performance for predicting pathological complete response (PCR) at two cycles of neoadjuvant chemotherapy (NAC) according to different tumour shrinkage patterns of luminal breast cancer and to assess the evaluative accuracy of ADC value combined with other clinicopathological indicators.

MATERIALS AND METHODS

Sixty-one patients who underwent NAC for histopathologically confirmed breast cancer were enrolled in this retrospective study. The ADC values of different shrinkage patterns (concentric shrinkage, nest or dendritic shrinkage, and mixed shrinkage) for tumours shown by diffusion-weighted imaging (DWI) were measured independently using three ROI placement methods (single-round, three-round, and freehand). Intraclass correlation coefficients (ICCs) were used to assess the interobserver variability in the ADC values. Multivariate logistic regression analysis was performed to identify the independent predictors of PCR.

RESULTS

The best placement method found was single-round ROI in all the patients (AUC=0.863). When analysed separately, the effectiveness results differed: the single-round method was optimal for concentrically shrinking tumours (AUC=0.970); the freehand method was optimal for nest or dendritically shrinking tumours (AUC=0.714); and the three-round method was optimal for mixed shrinking tumours (AUC=0.975). Multivariate logistic analysis showed that oestrogen receptor (ER), ΔADC% and tumour diameter reduction (ΔD%) were independent factors in evaluating the PCR.

CONCLUSION

The methods for measuring ADC values vary across different shrinkage patterns of luminal tumours. ΔADC%, ER and ΔD% were independent factors for evaluating the PCR.

Histogram analysis of ADC in rectal cancer: associations with different histopathological findings including expression of EGFR, Hif1-alpha, VEGF, p53, PD1, and KI 67. A preliminary study

Functional imaging modalities like Diffusion-weighted imaging are increasingly used to predict tumor behavior like cellularity and vascularity in different tumors. Histogram analysis is an emergent imaging analysis, in which every voxel is used to obtain a histogram and therefore statistically information about tumors can be provided. The purpose of this study was to elucidate possible associations between ADC histogram parameters and several immunhistochemical features in rectal cancer. Overall, 11 patients with histologically proven rectal cancer were included into the study. There were 2 (18.18%) females and 9 males with a mean age of 67.1 years. KI 67-index, expression of p53, EGFR, VEGF, and Hif1-alpha were semiautomatically estimated. The tumors were divided into PD1-positive and PD1-negative lesions. ADC histogram analysis was performed as a whole lesion measurement using an in-house matlab application. Spearman's correlation analysis revealed a strong correlation between EGFR expression and ADCmax (p=0.72, P=0.02). None of the vascular parameters (VEGF, Hif1-alpha) correlated with ADC parameters. Kurtosis and skewness correlated inversely with p53 expression (p=-0.64, P=0.03 and p=-0.81, P=0.002, respectively). ADCmedian and ADCmode correlated with Ki67 (p=-0.62, P=0.04 and p=-0.65, P=0.03, respectively). PD1-positive tumors showed statistically significant lower ADCmax values in comparison to PD1-negative tumors, 1.93 ± 0.36 vs 2.32 ± 0.47×10^-3mm²/s, p=0.04. Several associations were identified between histogram parameter derived from ADC maps and EGFR, KI 67 and p53 expression in rectal cancer. Furthermore, ADCmax was different between PD1 positive and PD1 negative tumors indicating an important role of ADC parameters for possible future treatment prediction.

Appearance Constrained Semi-Automatic Segmentation from DCE-MRI is Reproducible and Feasible for Breast Cancer Radiomics: A Feasibility Study

We present a segmentation approach that combines GrowCut (GC) with cancer-specific multi-parametric Gaussian Mixture Model (GCGMM) to produce accurate and reproducible segmentations. We evaluated GCGMM using a retrospectively collected 75 invasive ductal carcinoma with ERPR+ HER2− (n = 15), triple negative (TN) (n = 9), and ER-HER2+ (n = 57) cancers with variable presentation (mass and non-mass enhancement) and background parenchymal enhancement (mild and marked). Expert delineated manual contours were used to assess the segmentation performance using Dice coefficient (DSC), mean surface distance (mSD), Hausdorff distance, and volume ratio (VR). GCGMM segmentations were significantly more accurate than GrowCut (GC) and fuzzy c-means clustering (FCM). GCGMM’s segmentations and the texture features computed from those segmentations were the most reproducible compared with manual delineations and other analyzed segmentation methods. Finally, random forest (RF) classifier trained with leave-one-out cross-validation using features extracted from GCGMM segmentation resulted in the best accuracy for ER-HER2+ vs. ERPR+/TN (GCGMM 0.95, expert 0.95, GC 0.90, FCM 0.92) and for ERPR + HER2− vs. TN (GCGMM 0.92, expert 0.91, GC 0.77, FCM 0.83).