Intra-tumoural lipid composition and lymphovascular invasion in breast cancer via non-invasive magnetic resonance spectroscopy

Ultrasound-Based Deep Learning Radiomics Nomogram for the Assessment of Lymphovascular Invasion in Invasive Breast Cancer: A Multicenter Study

RATIONALE AND OBJECTIVES

The aim of this study was to develop a deep learning radiomics nomogram (DLRN) based on B-mode ultrasound (BMUS) and color doppler flow imaging (CDFI) images for preoperative assessment of lymphovascular invasion (LVI) status in invasive breast cancer (IBC).

MATERIALS AND METHODS

In this multicenter, retrospective study, 832 pathologically confirmed IBC patients were recruited from eight hospitals. The samples were divided into training, internal test, and external test sets. Deep learning and handcrafted radiomics features reflecting tumor phenotypes on BMUS and CDFI images were extracted. The BMUS score and CDFI score were calculated after radiomics feature selection. Subsequently, a DLRN was developed based on the scores and independent clinic-ultrasonic risk variables. The performance of the DLRN was evaluated for calibration, discrimination, and clinical usefulness.

RESULTS

The DLRN predicted the LVI with accuracy, achieving an area under the receiver operating characteristic curve of 0.93 (95% CI 0.90-0.95), 0.91 (95% CI 0.87-0.95), and 0.91 (95% CI 0.86-0.94) in the training, internal test, and external test sets, respectively, with good calibration. The DLRN demonstrated superior performance compared to the clinical model and single scores across all three sets (p < 0.05). Decision curve analysis and clinical impact curve confirmed the clinical utility of the model. Furthermore, significant enhancements in net reclassification improvement (NRI) and integrated discrimination improvement (IDI) indicated that the two scores could serve as highly valuable biomarkers for assessing LVI.

CONCLUSION

The DLRN exhibited strong predictive value for LVI in IBC, providing valuable information for individualized treatment decisions.

Spatial heterogeneity of peri-tumoural lipid composition in postmenopausal patients with oestrogen receptor positive breast cancer

Deregulation of lipid composition in adipose tissue adjacent to breast tumour is observed in ex vivo and animal models. Novel non-invasive magnetic resonance imaging (MRI) allows rapid lipid mapping of the human whole breast. We set out to elucidate the spatial heterogeneity of peri-tumoural lipid composition in postmenopausal patients with oestrogen receptor positive (ER +) breast cancer. Thirteen participants (mean age, 62 ± [SD] 6 years) with ER + breast cancer and 13 age-matched postmenopausal healthy controls were scanned on MRI. The number of double bonds in triglycerides was computed from MRI images to derive lipid composition maps of monounsaturated, polyunsaturated, and saturated fatty acids (MUFA, PUFA, SFA). The spatial heterogeneity measures (mean, median, skewness, entropy and kurtosis) of lipid composition in the peri-tumoural region and the whole breast of participants and in the whole breast of controls were computed. The Ki-67 proliferative activity marker and CD163 antibody on tumour-associated macrophages were assessed histologically. Mann Whitney U or Wilcoxon tests and Spearman's coefficients were used to assess group differences and correlations, respectively. For comparison against the whole breast in participants, peri-tumoural MUFA had a lower mean (median (IQR), 0.40 (0.02), p < .001), lower median (0.42 (0.02), p < .001), a negative skewness with lower magnitude (- 1.65 (0.77), p = .001), higher entropy (4.35 (0.64), p = .007) and lower kurtosis (5.13 (3.99), p = .001). Peri-tumoural PUFA had a lower mean (p < .001), lower median (p < .001), a positive skewness with higher magnitude (p = .005) and lower entropy (p = .002). Peri-tumoural SFA had a higher mean (p < .001), higher median (p < .001), a positive skewness with lower magnitude (p < .001) and lower entropy (p = .012). For comparison against the whole breast in controls, peri-tumoural MUFA had a negative skewness with lower magnitude (p = .01) and lower kurtosis (p = .009), however there was no difference in PUFA or SFA. CD163 moderately correlated with peri-tumoural MUFA skewness (rs = - .64), PUFA entropy (rs = .63) and SFA skewness (rs = .59). There was a lower MUFA and PUFA while a higher SFA, and a higher heterogeneity of MUFA while a lower heterogeneity of PUFA and SFA, in the peri-tumoural region in comparison with the whole breast tissue. The degree of lipid deregulation was associated with inflammation as indicated by CD163 antibody on macrophages, serving as potential marker for early diagnosis and response to therapy.

Ultrasound radiomics-based nomogram to predict lymphovascular invasion in invasive breast cancer: a multicenter, retrospective study

OBJECTIVES

To develop and validate an ultrasound (US) radiomics-based nomogram for the preoperative prediction of the lymphovascular invasion (LVI) status in patients with invasive breast cancer (IBC).

MATERIALS AND METHODS

In this multicentre, retrospective study, 456 consecutive women were enrolled from three institutions. Institutions 1 and 2 were used to train (n = 320) and test (n = 136), and 130 patients from institution 3 were used for external validation. Radiomics features that reflected tumour information were derived from grey-scale US images. The least absolute shrinkage and selection operator and the maximum relevance minimum redundancy (mRMR) algorithm were used for feature selection and radiomics signature (RS) building. US radiomics-based nomogram was constructed by using multivariable logistic regression analysis. Predictive performance was assessed with the receiving operating characteristic curve, discrimination, and calibration.

RESULTS

The nomogram based on clinico-ultrasonic features (menopausal status, US-reported lymph node status, posterior echo features) and RS yielded an optimal AUC of 0.88 (95% confidence interval [CI], 0.84-0.91), 0.89 (95% CI, 0.84-0.94) and 0.95 (95% CI, 0.92-0.99) in the training, internal and external validation cohort. The nomogram outperformed the clinico-ultrasonic and RS model (p < 0.05). The nomogram performed favourable discrimination (C-index, 0.88; 95% CI: 0.84-0.91) and was confirmed in the validation (0.88 for internal, 0.95 for external) cohorts. The calibration and decision curve demonstrated the nomogram showed good calibration and was clinically useful.

CONCLUSIONS

The radiomics nomogram incorporated in the RS and US and the clinical findings exhibited favourable preoperative individualised prediction of LVI.

CLINICAL RELEVANCE STATEMENT

The US radiomics-based nomogram incorporating menopausal status, posterior echo features, US reported-ALN status, and radiomics signature has the potential to predict lymphovascular invasion in patients with invasive breast cancer.

KEY POINTS

• The clinico-ultrsonic model of menopausal status, posterior echo features, and US-reported ALN status achieved a better predictive efficacy for LVI than either of them alone. • The radiomics nomogram showed optimal prediction in predicting LVI from patients with IBC (ROC, 0.88 and 0.89 in the training and validation sets). • A nomogram demonstrated favourable performance (area under the receiver operating characteristic curve, 0.95) and well calibration (C-index, 0.95) in an independent validation cohort (n = 130).

Towards detection of early response in neoadjuvant chemotherapy of breast cancer using Bayesian intravoxel incoherent motion

Introduction

The early identification of good responders to neoadjuvant chemotherapy (NACT) holds a significant potential in the optimal treatment of breast cancer. A recent Bayesian approach has been postulated to improve the accuracy of the intravoxel incoherent motion (IVIM) model for clinical translation. This study examined the prediction and early sensitivity of Bayesian IVIM to NACT response.

Materials and methods

Seventeen female patients with breast cancer were scanned at baseline and 16 patients were scanned after Cycle 1. Tissue diffusion and perfusion from Bayesian IVIM were calculated at baseline with percentage change at Cycle 1 computed with reference to baseline. Cellular proliferative activity marker Ki-67 was obtained semi-quantitatively with percentage change at excision computed with reference to core biopsy.

Results

The perfusion fraction showed a significant difference (p = 0.042) in percentage change between responder groups at Cycle 1, with a decrease in good responders [-7.98% (-19.47-1.73), n = 7] and an increase in poor responders [10.04% (5.09-28.93), n = 9]. There was a significant correlation between percentage change in perfusion fraction and percentage change in Ki-67 (p = 0.042). Tissue diffusion and pseudodiffusion showed no significant difference in percentage change between groups at Cycle 1, nor was there a significant correlation against percentage change in Ki-67. Perfusion fraction, tissue diffusion, and pseudodiffusion showed no significant difference between groups at baseline, nor was there a significant correlation against Ki-67 from core biopsy.

Conclusion

The alteration in tumour perfusion fraction from the Bayesian IVIM model, in association with cellular proliferation, showed early sensitivity to good responders in NACT.

Clinical trial registration

https://clinicaltrials.gov/ct2/show/NCT03501394, identifier NCT03501394.

Receptor Positive Breast Lesions and Status of Axillary Lymph Node

Objective This study aimed to examine the axillary pathological condition through the development of an assessment framework for hormone receptor-positive breast cancer subsequent to neoadjuvant chemotherapy (NAC). Furthermore, the primary objective of this study was to examine the association between axillary status and breast tumors that are positive for hormone receptors following neoadjuvant chemotherapy. Methodology The present retrospective investigation encompassed a cohort including 300 individuals who were administered neoadjuvant chemotherapy before receiving surgical intervention. The data collection period for this study was from September 2021 to December 2022. All patients received neoadjuvant chemotherapy, per the guidelines established by the National Comprehensive Cancer Network (NCCN). We divided patients into two distinct groups: a test set of 250 patients and a validity set of 50 patients. Patients with no evidence of lymphoid involvement underwent a biopsy of sentinel lymph nodes (SLNB) and would have undergone axillary dissection if the biopsy results had indicated positive findings. A logistic regression analysis was employed to investigate the variables linked to the presence of residual positive axillary lymph nodes in the test set. Subsequently, a multivariate analysis was conducted on the variables that exhibited a p-value below 0.2 in the univariate study. In addition, a value of 1 was assigned to all risk factors to construct a comprehensive correlation prediction model. Results The participants included in this study had a mean age and body mass index (BMI) of 46.24 ± 9.1 years and 25.8 ± 2.5 kg, respectively. The present investigation examined the presence of pathological axillary metastases in a cohort of 188 patients, which accounted for 62.55% of the total sample, by utilizing core-needle biopsy. Furthermore, the incidence rates of individuals presenting with clinical T1 were reported as 14.6%, while 55.2% cases of T2, 17.8% cases of T3, and 13% cases of T4 tumors were reported, respectively. Of the overall occurrences, the prevailing histological subtype was invasive ductal carcinoma, accounting for 91.4% (222 out of 243) of the cases. Multiple criteria were identified as independent predictors of the presence of residual positive axillary lymph nodes. The factors under consideration encompass lymphovascular invasion (odds ratio= 7.108; 95% breast cancer stage (odd ratio = 5.025; 95%, HER2 negativity (odd ratio= 2.997), low Ki-67 expression (odd ratio = 4.231), and suspected positive axillary lymph nodes before surgical intervention. Conclusion The present study presents a novel prediction model that integrates imaging and pathology data, aiming to assist patients and healthcare practitioners in evaluating the efficacy of NAC for hormone-receptor-positive breast tumors. The model holds particular significance for individuals who exhibit clinical positivity in their lymph nodes (LNs). Consequently, the model has the potential to provide guidance for the management of axillary lymph nodes and prevent unnecessary dissection.

Evaluation of Lymphatic Vessel Invasion Determined by D2-40 Using Preoperative MRI-Based Radiomics for Invasive Breast Cancer

RATIONALE AND OBJECTIVES

Preoperative prediction of LVI status can facilitate personalized therapeutic planning. This study aims to investigate the efficacy of preoperative MRI-based radiomics for predicting lymphatic vessel invasion (LVI) determined by D2-40 in patients with invasive breast cancer.

MATERIALS AND METHODS

A total of 203 patients with pathologically confirmed invasive breast cancer, who underwent preoperative breast MRI, were retrospectively enrolled and randomly assigned to the following cohorts: training cohort (n=141) and test cohort (n=62). Then, univariate and multivariate logistic regression were performed to select independent risk factors and build a clinical model. Afterwards, least absolute shrinkage and selection operator (LASSO) logistic regression was performed to select predictive features extracted from the early and delay enhancement dynamic contrast-enhanced (DCE)-MRI images, and a radiomics signature was established. Subsequently, a nomogram model was constructed by incorporating the radiomics score and risk factors. Receiver operating characteristic curves were performed to determine the performance of various models. The efficacy of the various models was evaluated using calibration and decision curves.

RESULTS

Fourteen radiomics features were selected to construct the radiomics model. The size of the lymph node was identified as an independent risk factor of the clinical model. The nomogram model demonstrated the best calibration and discrimination performance in both the training and test cohorts, with an area under the curve of 0.873 (95% confidence interval [CI]: 0.807-0.923) and 0.902 (95% CI: 0.800-0.963), respectively. The decision curve illustrated that the nomogram model added more net benefits, when compared to the radiomics signature and clinical model.

CONCLUSION

The nomogram model based on preoperative DCE-MRI images exhibits satisfactory efficacy for the noninvasive prediction of LVI determined by D2-40 in invasive breast cancer.

Using Machine Learning Methods to Assess Lymphovascular Invasion and Survival in Breast Cancer: Performance of Combining Preoperative Clinical and MRI Characteristics

BACKGROUND

Preoperative assessment of lymphovascular invasion (LVI) in invasive breast cancer (IBC) is of high clinical relevance for treatment decision-making and prognosis.

PURPOSE

To investigate the associations of preoperative clinical and magnetic resonance imaging (MRI) characteristics with LVI and disease-free survival (DFS) by using machine learning methods in patients with IBC.

STUDY TYPE

Retrospective.

POPULATION

Five hundred and seventy-five women (range: 24-79 years) with IBC who underwent preoperative MRI examinations at two hospitals, divided into the training (N = 386) and validation datasets (N = 189).

FIELD STRENGTH/SEQUENCE

Axial fat-suppressed T2-weighted turbo spin-echo sequence and dynamic contrast-enhanced with fat-suppressed T1-weighted three-dimensional gradient echo imaging.

ASSESSMENT

MRI characteristics (clinical T stage, breast edema score, MRI axillary lymph node status, multicentricity or multifocality, enhancement pattern, adjacent vessel sign, and increased ipsilateral vascularity) were reviewed independently by three radiologists. Logistic regression (LR), eXtreme Gradient Boosting (XGBoost), k-Nearest Neighbor (KNN), and Support Vector Machine (SVM) algorithms were used to establish the models by combing preoperative clinical and MRI characteristics for assessing LVI status in the training dataset, and the methods were further applied in the validation dataset. The LVI score was calculated using the best-performing of the four models to analyze the association with DFS.

STATISTICAL TESTS

Chi-squared tests, variance inflation factors, receiver operating characteristics (ROC), Kaplan-Meier curve, log-rank, Cox regression, and intraclass correlation coefficient were performed. The area under the ROC curve (AUC) and hazard ratios (HR) were calculated. A P-value <0.05 was considered statistically significant.

RESULTS

The model established by the XGBoost algorithm had better performance than LR, SVM, and KNN models, achieving an AUC of 0.832 (95% confidence interval [CI]: 0.789, 0.876) in the training dataset and 0.838 (95% CI: 0.775, 0.901) in the validation dataset. The LVI score established by the XGBoost model was an independent indicator of DFS (adjusted HR: 2.66, 95% CI: 1.22-5.80).

DATA CONCLUSION

The XGBoost model based on preoperative clinical and MRI characteristics may help to investigate the LVI status and survival in patients with IBC.

LEVEL OF EVIDENCE

3 TECHNICAL EFFICACY: Stage 2.

Hyperpolarised 13C-MRI using 13C-pyruvate in breast cancer: A review

Tumour metabolism can be imaged with a novel imaging technique termed hyperpolarised carbon-13 (13C)-MRI using probes, i.e., endogenously found molecules that are labeled with 13C. Hyperpolarisation of the 13C label increases the sensitivity to a level that allows dynamic imaging of the distribution and metabolism of the probes. Dynamic imaging of [1-13C]pyruvate metabolism is of particular biological interest in cancer because of the Warburg effect resulting in the intratumoural accumulation of [1-13C]pyruvate and conversion to [1-13C]lactate. Numerous preclinical studies in breast cancer and other tumours have shown that hyperpolarised 13C-pyruvate has potential for metabolic phenotyping and response assessment at earlier timepoints than the current clinical imaging techniques allow. The clinical feasibility of hyperpolarised 13C-MRI after the injection of pyruvate in patients with breast cancer has now been demonstrated, with increased 13C-label exchange between pyruvate and lactate present in higher grade tumours with associated increased expression of the monocarboxylate transporter 1 (MCT1), the transmembrane transporter mediating intracellular pyruvate uptake, and lactate dehydrogenase (LDH) as the enzyme catalysing the conversion of pyruvate to lactate. Furthermore, a study in patients with breast cancer undergoing neoadjuvant chemotherapy suggested that early changes in 13C-label exchange can distinguish between patients who reach pathologic complete response (pCR) and those who do not. This review summarises the current literature on preclinical and clinical research on hyperpolarised 13C-MRI with [1-13C]-pyruvate in breast cancer imaging.

Unveiling the MUFA-Cancer Connection: Insights from Endogenous and Exogenous Perspectives

Monounsaturated fatty acids (MUFAs) have been the subject of extensive research in the field of cancer due to their potential role in its prevention and treatment. MUFAs can be consumed through the diet or endogenously biosynthesized. Stearoyl-CoA desaturases (SCDs) are key enzymes involved in the endogenous synthesis of MUFAs, and their expression and activity have been found to be increased in various types of cancer. In addition, diets rich in MUFAs have been associated with cancer risk in epidemiological studies for certain types of carcinomas. This review provides an overview of the state-of-the-art literature on the associations between MUFA metabolism and cancer development and progression from human, animal, and cellular studies. We discuss the impact of MUFAs on cancer development, including their effects on cancer cell growth, migration, survival, and cell signaling pathways, to provide new insights on the role of MUFAs in cancer biology.

Probing lipids relaxation times in breast cancer using magnetic resonance spectroscopic fingerprinting

OBJECTIVES

To investigate the clinical relevance of the relaxation times of lipids within breast cancer and normal fibroglandular tissue in vivo, using magnetic resonance spectroscopic fingerprinting (MRSF).

METHODS

Twelve patients with biopsy-confirmed breast cancer and 14 healthy controls were prospectively scanned at 3 T using a protocol consisting of diffusion tensor imaging (DTI), MRSF, and dynamic contrast-enhanced (DCE) MRI. Single-voxel MRSF data was recorded from the tumor (patients) - identified using DTI - or normal fibroglandular tissue (controls), in under 20 s. MRSF data was analyzed using in-house software. Linear mixed model analysis was used to compare the relaxation times of lipids in breast cancer VOIs vs. normal fibroglandular tissue.

RESULTS

Seven distinguished lipid metabolite peaks were identified and their relaxation times were recorded. Of them, several exhibited statistically significant changes between controls and patients, with strong significance (p < 10^-3) recorded for several of the lipid resonances at 1.3 ppm (T₁ = 355 ± 17 ms vs. 389 ± 27 ms), 4.1 ppm (T₁ = 255 ± 86 ms vs. 127 ± 33 ms), 5.22 ppm (T₁ = 724 ± 81 ms vs. 516 ± 62 ms), and 5.31 ppm (T₂ = 56 ± 5 ms vs. 44 ± 3.5 ms, respectively).

CONCLUSIONS

The application of MRSF to breast cancer imaging is feasible and achievable in clinically relevant scan time. Further studies are required to verify and comprehend the underling biological mechanism behind the differences in lipid relaxation times in cancer and normal fibroglandular tissue.

KEY POINTS

•The relaxation times of lipids in breast tissue are potential markers for quantitative characterization of the normal fibroglandular tissue and cancer. •Lipid relaxation times can be acquired rapidly in a clinically relevant manner using a single-voxel technique, termed MRSF. •Relaxation times of T₁ at 1.3 ppm, 4.1 ppm, and 5.22 ppm, as well as of T₂ at 5.31 ppm, were significantly different between measurements within breast cancer and the normal fibroglandular tissue.

Overview of Artificial Intelligence in Breast Cancer Medical Imaging

The heavy global burden and mortality of breast cancer emphasize the importance of early diagnosis and treatment. Imaging detection is one of the main tools used in clinical practice for screening, diagnosis, and treatment efficacy evaluation, and can visualize changes in tumor size and texture before and after treatment. The overwhelming number of images, which lead to a heavy workload for radiologists and a sluggish reporting period, suggests the need for computer-aid detection techniques and platform. In addition, complex and changeable image features, heterogeneous quality of images, and inconsistent interpretation by different radiologists and medical institutions constitute the primary difficulties in breast cancer screening and imaging diagnosis. The advancement of imaging-based artificial intelligence (AI)-assisted tumor diagnosis is an ideal strategy for improving imaging diagnosis efficient and accuracy. By learning from image data input and constructing algorithm models, AI is able to recognize, segment, and diagnose tumor lesion automatically, showing promising application prospects. Furthermore, the rapid advancement of "omics" promotes a deeper and more comprehensive recognition of the nature of cancer. The fascinating relationship between tumor image and molecular characteristics has attracted attention to the radiomic and radiogenomics, which allow us to perform analysis and detection on the molecular level with no need for invasive operations. In this review, we integrate the current developments in AI-assisted imaging diagnosis and discuss the advances of AI-based breast cancer precise diagnosis from a clinical point of view. Although AI-assisted imaging breast cancer screening and detection is an emerging field and draws much attention, the clinical application of AI in tumor lesion recognition, segmentation, and diagnosis is still limited to research or in limited patients' cohort. Randomized clinical trials based on large and high-quality cohort are lacking. This review aims to describe the progress of the imaging-based AI application in breast cancer screening and diagnosis for clinicians.

Predicting Axillary Response in Hormone Receptor-Positive Breast Cancer after Neoadjuvant Chemotherapy Using Real-World Data

Purpose

To develop a scoring system for hormone receptor-positive (HR+) breast cancer patients who are expected to achieve axillary pathological complete response (pCR) after neoadjuvant chemotherapy (NAC). To confirm the correlation between axillary status and survival rate in HR+ breast cancer after NAC.

Methods

Women from the Shanghai Jiao Tong University Breast Cancer Database (SJTU-BCDB) who underwent NAC for cT1-4N1-3M0 primary HR+ breast cancer between 2009 and 2018 were included in the study. In this case, patient follow up was performed until 2022 for those with complete data before and after NAC. The main outcome measures were the axillary pCR rate, overall survival (OS), and disease-free survival (DFS). The patients were randomly assigned to a test set (n = 175) and a validation set (n = 68) in a 7 : 3 ratio. A prediction risk score was then developed based on the odds ratios from the multivariate analysis of the test set (n = 175) before being validated in the validation set (n = 68). Finally, the Kaplan-Meier curves were used to explore the survival on this score system.

Results

From the database, 243 women were included, and the median follow-up period was 47.5 months (95% confidence interval: 41.9-53.1). The axillary pCR rate was 18.9% (46 of 243), with the independent predictors of residual positive axillary lymph nodes (LNs) being lymphovascular invasion (LVI), breast conserving surgery (BCS), Ki67 < 14%, HER2 negativity, positive lymph nodes in ultrasound (US) before surgery, and stage III histological grade (All, P < 0.05). Using the above predictors of the model, the receiver operating characteristic (ROC) curve was used for calibration and inspection, with values for the test and validation sets being 0.847 (P < 0.001; 95% CI: 0.769, 0.925) and 0.813 (P < 0.001; 95% CI: 0.741, 0.885), respectively. The total risk score ranged from 0 to 6 for the multivariate analysis, and from this range, a risk score of 0-2 was defined as a low-risk group, while scores of 3-6 were defined as the high-risk one. By constructing the survival curve, it was found that the 5-year OS rates for the low-risk and high-risk groups were 89.0% and 84.2% (P = 0.236). Similarly, the 5-year DFS rates for the low-risk and high-risk groups were 80% and 68.5% (P = 0.048), respectively. In addition, axillary pathological stages were significantly correlated with the overall survival (OS) and disease-free survival (DFS) (All, P < 0.05).

Conclusion

The prediction model showed good performance for HR + breast cancer. LVI, BCS, low Ki-67, HER2 negativity, suspected positive LNs before surgery, and stage III histological grade were all risk factors for residual positive axillary LNs. However, unlike pathological stages, achieving pCR in the axillary LNs does not affect the survival status.

Preoperative prediction of lymphovascular invasion in patients with T1 breast invasive ductal carcinoma based on radiomics nomogram using grayscale ultrasound

Purpose

The aim of this study was to develop a radiomics nomogram based on grayscale ultrasound (US) for preoperatively predicting Lymphovascular invasion (LVI) in patients with pathologically confirmed T1 (pT1) breast invasive ductal carcinoma (IDC).

Methods

One hundred and ninety-two patients with pT1 IDC between September 2020 and August 2022 were analyzed retrospectively. Study population was randomly divided in a 7: 3 ratio into a training dataset of 134 patients (37 patients with LVI-positive) and a validation dataset of 58 patients (19 patients with LVI-positive). Clinical information and conventional US (CUS) features (called clinic_CUS features) were recorded and evaluated to predict LVI. In the training dataset, independent predictors of clinic_CUS features were obtained by univariate and multivariate logistic regression analyses and incorporated into a clinic_CUS prediction model. In addition, radiomics features were extracted from the grayscale US images, and the radiomics score (Radscore) was constructed after radiomics feature selection. Subsequent multivariate logistic regression analysis was also performed for Radscore and the independent predictors of clinic_CUS features, and a radiomics nomogram was developed. The performance of the nomogram model was evaluated via its discrimination, calibration, and clinical usefulness.

Results

The US reported axillary lymph node metastasis (LNM) (US_LNM) status and tumor margin were determined as independent risk factors, which were combined for the construction of clinic_CUS prediction model for LVI in pT1 IDC. Moreover, tumor margin, US_LNM status and Radscore were independent predictors, incorporated as the radiomics nomogram model, which achieved a superior discrimination to the clinic_CUS model in the training dataset (AUC: 0.849 vs. 0.747; P < 0.001) and validation dataset (AUC: 0.854 vs. 0.713; P = 0.001). Calibration curve for the radiomic nomogram showed good concordance between predicted and actual probability. Furthermore, decision curve analysis (DCA) confirmed that the radiomics nomogram had higher clinical net benefit than the clinic_CUS model.

Conclusion

The US-based radiomics nomogram, incorporating tumor margin, US_LNM status and Radscore, showed a satisfactory preoperative prediction of LVI in pT1 IDC patients.

Peri-tumoural spatial distribution of lipid composition and tubule formation in breast cancer

BACKGROUND

Response guided treatment in breast cancer is highly desirable, but the effectiveness is only established based on residual cellularity from histopathological analysis after surgery. Tubule formation, a key component of grading score, is directly associated with cellularity, with significant implications on prognosis. Peri-tumoural lipid composition, a potential marker, can be rapidly mapped across the entire breast using novel method of chemical shift-encoded imaging, enabling the quantification of spatial distribution. We hypothesise that peri-tumoural spatial distribution of lipid composition is sensitive to tumour cellular differentiation and proliferative activity.

METHODS

Twenty whole tumour specimens freshly excised from patients with invasive ductal carcinoma (9 Score 2 and 11 Score 3 in tubule formation) were scanned on a 3 T clinical scanner (Achieva TX, Philips Healthcare). Quantitative lipid composition maps were acquired for polyunsaturated, monounsaturated, and saturated fatty acids (PUFA, MUFA, SFA). The peri-tumoural spatial distribution (mean, skewness, entropy and kurtosis) of each lipid constituent were then computed. The proliferative activity marker Ki-67 and tumour-infiltrating lymphocytes (TILs) were assessed histologically.

RESULTS

For MUFA, there were significant differences between groups in mean (p = 0.0119), skewness (p = 0.0116), entropy (p = 0.0223), kurtosis (p = 0.0381), and correlations against Ki-67 in mean (ρ = -0.5414), skewness (ρ = 0.6045) and entropy (ρ = 0.6677), and TILs in mean (ρ = -0.4621). For SFA, there were significant differences between groups in mean (p = 0.0329) and skewness (p = 0.0111), and correlation against Ki-67 in mean (ρ = 0.5910). For PUFA, there was no significant difference in mean, skewness, entropy or kurtosis between the groups.

CONCLUSIONS

There was an association between peri-tumoural spatial distribution of lipid composition with tumour cellular differentiation and proliferation. Peri-tumoural lipid composition imaging might have potential in non-invasive quantitative assessment of patients with breast cancer for treatment planning and monitoring.

MRI Radiomics of Breast Cancer: Machine Learning-Based Prediction of Lymphovascular Invasion Status

RATIONALE AND OBJECTIVES

In patients with breast cancer (BC), lymphovascular invasion (LVI) status is considered an important prognostic factor. We aimed to develop machine learning (ML)-based radiomics models for the prediction of LVI status in patients with BC, using preoperative MRI images.

MATERIALS AND METHODS

This retrospective study included patients with BC with known LVI status and preoperative MRI. The dataset was split into training and unseen testing sets by stratified sampling with a 2:1 ratio. 2D and 3D radiomic features were extracted from contrast-enhanced T1 weighted images (C+T1W) and apparent diffusion coefficient (ADC) maps. The reliability of the features was assessed with two radiologists' segmentation data. Dimension reduction was done with reliability analysis, multi-collinearity analysis, removal of low-variance features, and feature selection. ML models were created with base, tuned, and boosted random forest algorithms.

RESULT

A total of 128 lesions (LVI-positive, 76; LVI-negative, 52) were included. The best model performance was achieved with tunning and boosting model based on 3D ADC maps and selected four radiomic features. The area under the curve and accuracy were 0.726 and 63.5% in the training data, 0.732 and 76.7% in the test data, respectively. The overall sensitivity and positive predictive values were 68% and 69.6% in the training data, 84.6% and 78.6% in the test data, respectively.

CONCLUSION

ML and radiomics based on 3D segmentation of ADC maps can be used to predict LVI status in BC, with satisfying performance.

Lactate Dehydrogenase-A (LDH-A) Preserves Cancer Stemness and Recruitment of Tumor-Associated Macrophages to Promote Breast Cancer Progression

Increasing evidence reveals that breast cancer stem cells (BCSCs) subtypes with distinct properties are regulated by their abnormal metabolic changes; however, the specific molecular mechanism and its relationship with tumor microenvironment (TME) are not clear. In this study, we explored the mechanism of lactate dehydrogenase A (LDHA), a crucial glycolytic enzyme, in maintaining cancer stemness and BCSCs plasticity, and promoting the interaction of BCSCs with tumor associated macrophages (TAMs). Firstly, the expression of LDHA in breast cancer tissues was much higher than that in adjacent tissues and correlated with the clinical progression and prognosis of breast cancer patients based on The Cancer Genome Atlas (TCGA) data set. Moreover, the orthotopic tumor growth and pulmonary metastasis were remarkable inhibited in mice inoculated with 4T1-shLdha cells. Secondly, the properties of cancer stemness were significantly suppressed in MDA-MB-231-shLDHA or A549-shLDHA cancer cells, including the decrease of ALDH⁺ cells proportion, the repression of sphere formation and cellular migration, and the reduction of stemness genes (SOX2, OCT4, and NANOG) expression. However, the proportion of ALDH⁺ cells (epithelial-like BCSCs, E-BCSCs) was increased and the proportion of CD44⁺ CD24^- cells (mesenchyme-like BCSCs, M-BCSCs) was decreased after LDHA silencing, suggesting a regulatory role of LDHA in E-BCSCs/M-BCSCs transformation in mouse breast cancer cells. Thirdly, the expression of epithelial marker E-cadherin, proved to interact with LDHA, was obviously increased in LDHA-silencing cancer cells. The recruitment of TAMs and the secretion of CCL2 were dramatically reduced after LDHA was knocked down in vitro and in vivo. Taken together, LDHA mediates a vicious cycle of mutual promotion between BCSCs plasticity and TAMs infiltration, which may provide an effective treatment strategy by targeting LDHA for breast cancer patients.