Establishment and Verification of a Bagged-Trees-Based Model for Prediction of Sentinel Lymph Node Metastasis for Early Breast Cancer Patients

A novel machine learning prediction model for metastasis in breast cancer

BACKGROUND

Breast cancer (BC) metastasis is the common cause of high mortality. Conventional prognostic criteria cannot accurately predict the BC metastasis risk. The machine learning technologies can overcome the disadvantage of conventional models.

AIM

We developed a model to predict BC metastasis using the random survival forest (RSF) method.

METHODS

Based on demographic data and routine clinical data, we used RSF-recursive feature elimination to identify the predictive variables and developed a model to predict metastasis using RSF method. The area under the receiver operating characteristic curve (AUROC) and Kaplan-Meier survival (KM) analyses were plotted to validate the predictive effect when C-index was plotted to assess the discrimination and Brier scores was plotted to assess the calibration of the predictive model.

RESULTS

We developed a metastasis prediction model comprising three variables (pathological stage, aspartate aminotransferase, and neutrophil count) selected by RSF-recursive feature elimination. The model was reliable and stable when assessed by the AUROC (0.932 in training set and 0.905 in validation set) and KM survival analyses (p < .0001). The C-indexes (0.959) and Brier score (0.097) also validated the good predictive ability of this model.

CONCLUSIONS

This model relies on routine data and examination indicators in real-time clinical practice and exhibits an accurate prediction performance without increasing the cost for patients. Using this model, clinicians can facilitate risk communication and provide precise and efficient individualized therapy to patients with breast cancer.

Prediction of sentinel lymph node metastasis in breast cancer patients based on preoperative features: a deep machine learning approach

Sentinel lymph node (SLN) biopsy is the standard surgical approach to detect lymph node metastasis in breast cancer. Machine learning is a novel tool that provides better accuracy for predicting positive SLN involvement in breast cancer patients. This study obtained data from 2890 surgical cases of breast cancer patients from two referral hospitals in Iran from 2000 to 2021. Patients whose SLN involvement status was identified were included in our study. The dataset consisted of preoperative features, including patient features, gestational factors, laboratory data, and tumoral features. In this study, TabNet, an end-to-end deep learning model, was proposed to predict SLN involvement in breast cancer patients. We compared the accuracy of our model with results from logistic regression analysis. A total of 1832 patients with an average age of 51 ± 12 years were included in our study, of which 697 (25.5%) had SLN involvement. On average, the TabNet model achieved an accuracy of 75%, precision of 81%, specificity of 70%, sensitivity of 87%, and AUC of 0.74, while the logistic model demonstrated an accuracy of 70%, precision of 73%, specificity of 65%, sensitivity of 79%, F1 score of 73%, and AUC of 0.70 in predicting the SLN involvement in patients. Vascular invasion, tumor size, core needle biopsy pathology, age, and FH had the most contributions to the TabNet model. The TabNet model outperformed the logistic regression model in all metrics, indicating that it is more effective in predicting SLN involvement in breast cancer patients based on preoperative data.

Prediction Model of Ocular Metastases in Gastric Adenocarcinoma: Machine Learning-Based Development and Interpretation Study

Background: Although gastric adenocarcinoma (GA) related ocular metastasis (OM) is rare, its occurrence indicates a more severe disease. We aimed to utilize machine learning (ML) to analyze the risk factors of GA-related OM and predict its risks. Methods: This is a retrospective cohort study. The clinical data of 3532 GA patients were collected and randomly classified into training and validation sets in a ratio of 7:3. Those with or without OM were classified into OM and non-OM (NOM) groups. Univariate and multivariate logistic regression analyses and least absolute shrinkage and selection operator were conducted. We integrated the variables identified through feature importance ranking and further refined the selection process using forward sequential feature selection based on random forest (RF) algorithm before incorporating them into the ML model. We applied six ML algorithms to construct the predictive GA model. The area under the receiver operating characteristic (ROC) curve indicated the model's predictive ability. Also, we established a network risk calculator based on the best performance model. We used Shapley additive interpretation (SHAP) to identify risk factors and to confirm the interpretability of the black box model. We have de-identified all patient details. Results: The ML model, consisting of 13 variables, achieved an optimal predictive performance using the gradient boosting machine (GBM) model, with an impressive area under the curve (AUC) of 0.997 in the test set. Utilizing the SHAP method, we identified crucial factors for OM in GA patients, including LDL, CA724, CEA, AFP, CA125, Hb, CA153, and Ca2+. Additionally, we validated the model's reliability through an analysis of two patient cases and developed a functional online web prediction calculator based on the GBM model. Conclusion: We used the ML method to establish a risk prediction model for GA-related OM and showed that GBM performed best among the six ML models. The model may identify patients with GA-related OM to provide early and timely treatment.

Pathologic complete response prediction in breast cancer lesion segmentation and neoadjuvant therapy

Objectives

Predicting whether axillary lymph nodes could achieve pathologic Complete Response (pCR) after breast cancer patients receive neoadjuvant chemotherapy helps make a quick follow-up treatment plan. This paper presents a novel method to achieve this prediction with the most effective medical imaging method, Dynamic Contrast-enhanced Magnetic Resonance Imaging (DCE-MRI).

Methods

In order to get an accurate prediction, we first proposed a two-step lesion segmentation method to extract the breast cancer lesion region from DCE-MRI images. With the segmented breast cancer lesion region, we then used a multi-modal fusion model to predict the probability of axillary lymph nodes achieving pCR.

Results

We collected 361 breast cancer samples from two hospitals to train and test the proposed segmentation model and the multi-modal fusion model. Both segmentation and prediction models obtained high accuracy.

Conclusion

The results show that our method is effective in both the segmentation task and the pCR prediction task. It suggests that the presented methods, especially the multi-modal fusion model, can be used for the prediction of treatment response in breast cancer, given data from noninvasive methods only.

An Australian-based nomogram predicting nodal positivity for breast cancer in older patients

INTRODUCTION

Regional nodal status is one of the most powerful prognosticators in breast cancer. The sentinel lymph node biopsy (SLNB) evaluates the first node in the axillary lymphatic basin theorised to drain the anatomical region of breast cancer. Recent literature has appropriately raised the query of the need for SLNB for breast cancer in older patients (BCOP). Though some early-stage older patients may safely have SLNB omitted, we are potentially missing the under-represented aggressive cancers. No sentinel lymph node metastases nomogram has been developed solely from BCOP data to date. This study aimed to identify older patients with breast cancer at risk for nodal involvement using a nomogram developed from their data alone.

METHODS

A retrospective analysis of prospectively collected data on BCOP (aged ≥70 years) was performed using the Breast Surgery Quality Audit (BQA). Inclusion criteria were T1-2 invasive breast cancer patients that underwent a SLNB from 1st January 2001 to 31st December 2019. The primary outcome was nodal involvement. Data obtained from the dataset included: age, tumour type, tumour size in millimeters, histological grade, lymphovascular invasion, oestrogen receptor status, progesterone receptor status, HER2 status and referral source. Binary logistic regression was used to develop a nomogram. The model was internally validated by splitting the data set (80% for training and 20% for testing). A receiver operating characteristic curve was developed, with the area under the curve (AUC) and a calibration plot.

RESULTS

There were 22,313 patients of which 14,856 (66.6%) were symptomatic presentations and 7457 (33.4%) were screen-detected. Invasive tumour type, tumour size, tumour grade, lymphovascular invasion, oestrogen receptors, and referral source indicated a statistically significant effect on predicting a nodal positivity event (Table 1). The AUC was 0.782 (95% CI 0.776-0.789) (Fig. 1a) and demonstrated good calibration (Fig. 1b). The negative predictive value established was 85%.

CONCLUSION

We have developed an Australian sentinel lymph node metastasis nomogram for BCOP using routine histopathological data obtained pre-operatively (Fig. 2). This is the first Australian nomogram, as well as the first nomogram developed solely for BCOP - and maintains a superior AUC compared to other well-established nomograms.

Multi-center study on predicting breast cancer lymph node status from core needle biopsy specimens using multi-modal and multi-instance deep learning

The objective of our study is to develop a deep learning model based on clinicopathological data and digital pathological image of core needle biopsy specimens for predicting breast cancer lymph node metastasis. We collected 3701 patients from the Fourth Hospital of Hebei Medical University and 190 patients from four medical centers in Hebei Province. Integrating clinicopathological data and image features build multi-modal and multi-instance (MMMI) deep learning model to obtain the final prediction. For predicting with or without lymph node metastasis, the AUC was 0.770, 0.709, 0.809 based on the clinicopathological features, WSI and MMMI, respectively. For predicting four classification of lymph node status (no metastasis, isolated tumor cells (ITCs), micrometastasis, and macrometastasis), the prediction based on clinicopathological features, WSI and MMMI were compared. The AUC for no metastasis was 0.770, 0.709, 0.809, respectively; ITCs were 0.619, 0.531, 0.634, respectively; micrometastasis were 0.636, 0.617, 0.691, respectively; and macrometastasis were 0.748, 0.691, 0.758, respectively. The MMMI model achieved the highest prediction accuracy. For prediction of different molecular types of breast cancer, MMMI demonstrated a better prediction accuracy for any type of lymph node status, especially in the molecular type of triple negative breast cancer (TNBC). In the external validation sets, MMMI also showed better prediction accuracy in the four classification, with AUC of 0.725, 0.757, 0.525, and 0.708, respectively. Finally, we developed a breast cancer lymph node metastasis prediction model based on a MMMI model. Through all cases tests, the results showed that the overall prediction ability was high.

Models for Predicting Sentinel and Non-sentinel Lymph Nodes Based on Pre-operative Ultrasonic Breast Imaging to Optimize Axillary Strategies

Axillary strategy decisions have become more complex and controversial in considering minimally traumatic therapy instead of sentinel lymph node biopsy, axillary lymph node dissection or regional nodal irradiation for people with breast cancer. The purpose of this study was to noninvasively predict sentinel lymph node (SLN) and non-sentinel lymph node (NSLN) status based on pre-operative sonographic and clinicopathologic features to determine optimal decisions regarding axillary therapy. In total, 701 patients with breast cancer from two independent centers were retrospectively analyzed. The SLN model (SLNM) for predicting SLN status and the NSLN model (NSLNM) for predicting NSLN status were trained based on a training set using the random-forest algorithm, and their performance was validated using an independent external test set. A receiver operating characteristic curve was drawn to obtain the area under the curve, which was used to assess performance. The area under the curve for the SLNM in the training and test, respectively, was 94.2% and 83.0%, and for the NSLNM, 99.5% and 92.7%. The SLNM and NSLNM accurately predicted that 61.46% (319/519) and 17.53% (91/519), respectively, of our participants were non-metastatic. The overall benefit of the three models was 78.99% in our participants. The two models for predicting SLN and NSLN status showed excellent application potential in optimizing axillary strategies.

"Real-world" radiomics from multi-vendor MRI: an original retrospective study on the prediction of nodal status and disease survival in breast cancer, as an exemplar to promote discussion of the wider issues

BACKGROUND

Most MRI radiomics studies to date, even multi-centre ones, have used "pure" datasets deliberately accrued from single-vendor, single-field-strength scanners. This does not reflect aspirations for the ultimate generalisability of AI models. We therefore investigated the development of a radiomics signature from heterogeneous data originating on six different imaging platforms, for a breast cancer exemplar, in order to provide input into future discussions of the viability of radiomics in "real-world" scenarios where image data are not controlled by specific trial protocols but reflective of routine clinical practice.

METHODS

One hundred fifty-six patients with pathologically proven breast cancer underwent multi-contrast MRI prior to neoadjuvant chemotherapy and/or surgery. From these, 92 patients were identified for whom T2-weighted, diffusion-weighted and contrast-enhanced T1-weighted sequences were available, as well as key clinicopathological variables. Regions-of-interest were drawn on the above image types and, from these, semantic and calculated radiomics features were derived. Classification models using a variety of methods, both with and without recursive feature elimination, were developed to predict pathological nodal status. Separately, we applied the same methods to analyse the information carried by the radiomic features regarding the originating scanner type and field strength. Repeated, ten-fold cross-validation was employed to verify the results. In parallel work, survival modelling was performed using random survival forests.

RESULTS

Prediction of nodal status yielded mean cross-validated AUC values of 0.735 ± 0.15 (SD) for clinical variables alone, 0.673 ± 0.16 (SD) for radiomic features only, and 0.764 ± 0.16 (SD) for radiomics and clinical features together. Prediction of scanner platform from the radiomics features yielded extremely high values of AUC between 0.91 and 1 for the different classes examined indicating the presence of confounding features for the nodal status classification task. Survival analysis, gave out-of-bag prediction errors of 19.3% (clinical features only), 36.9-51.8% (radiomic features from different combinations of image contrasts), and 26.7-35.6% (clinical plus radiomics features).

CONCLUSIONS

Radiomic classification models whose predictive ability was consistent with previous single-vendor, single-field strength studies have been obtained from multi-vendor, multi-field-strength data, despite clear confounding information being present. However, our sample size was too small to obtain useful survival modelling results.

Establishment of a model for predicting sentinel lymph node metastasis in early breast cancer based on contrast-enhanced ultrasound and clinicopathological features

Background

Sentinel lymph node (SLN) biopsy (SLNB) is the standard procedure for axillary staging in clinically node-negative (cN0) breast cancer patients. However, the positive rate of SLNs among cN0 stage patients is 26-35%. The identification of appropriate candidates for SLNB is quite challenging. This study aimed to establish and verify a predictive model of SLN metastasis using contrast-enhanced ultrasound (CEUS) and other clinicopathological indicators.

Methods

The clinicopathological data of 224 patients who had undergone SLNB at the Affiliated Cancer Hospital of Zhengzhou University from June 2018 to July 2019 were analyzed retrospectively. The risk prediction model of SLN metastasis was established by logistic regression analysis. According to the β value of each variable in the model, a risk score system of SLN metastasis was established and verified using the internal population. The predictive model was prospectively applied to 73 patients from July 2019 to September 2019 to evaluate the clinical value of the model in patients with early breast cancer.

Results

Multivariate analysis confirmed that body mass index (BMI), SLN aspect ratio of CEUS mode, SLN aspect ratio of mammography, lympho-vascular invasion, and cytokeratin (CK)5/6 were independent risk factors for SLN metastasis. A scoring system was established according to the above risk factors, and a receiver operating characteristic (ROC) curve was drawn. After internal- and external verification, a corrected ROC curve was drawn, respectively. The ROC curve of the modeling group, internal verification group, and external verification group was 0.9075 (95% CI: 0.8616-0.9534), 0.8766 (95% CI: 0.8192-0.9341), and 0.8505 (95% CI: 0.7333-0.9676), respectively.

Conclusions

We constructed and verified a prediction model of SLN metastasis in early breast cancer. The model has a specific predictive value for preoperative evaluation of SLN status.

Preoperative Nomogram for Predicting Sentinel Lymph Node Metastasis Risk in Breast Cancer: A Potential Application on Omitting Sentinel Lymph Node Biopsy

Background

Sentinel lymph node (SLN) biopsy is feasible for breast cancer (BC) patients with clinically negative axillary lymph nodes; however, complications develop in some patients after surgery, although SLN metastasis is rarely found. Previous predictive models contained parameters that relied on postoperative data, thus limiting their application in the preoperative setting. Therefore, it is necessary to find a new model for preoperative risk prediction for SLN metastasis to help clinicians facilitate individualized clinical decisions.

Materials and Methods

BC patients who underwent SLN biopsy in two different institutions were included in the training and validation cohorts. Demographic characteristics, preoperative tumor pathological features, and ultrasound findings were evaluated. Multivariate logistic regression was used to develop the nomogram. The discrimination, accuracy, and clinical usefulness of the nomogram were assessed using Harrell’s C-statistic and ROC analysis, the calibration curve, and the decision curve analysis, respectively.

Results

A total of 624 patients who met the inclusion criteria were enrolled, including 444 in the training cohort and 180 in the validation cohort. Young age, high BMI, high Ki67, large tumor size, indistinct tumor margins, calcifications, and an aspect ratio ≥1 were independent predictive factors for SLN metastasis of BC. Incorporating these parameters, the nomogram achieved a robust predictive performance with a C-index and accuracy of 0.92 and 0.85, and 0.82 and 0.80 in the training and validation cohorts, respectively. The calibration curves also fit well, and the decision curve analysis revealed that the nomogram was clinically useful.

Conclusions

We established a nomogram to preoperatively predict the risk of SLN metastasis in BC patients, providing a non-invasive approach in clinical practice and serving as a potential tool to identify BC patients who may omit unnecessary SLN biopsy.

Nomograms for Predicting Axillary Lymph Node Status Reconciled With Preoperative Breast Ultrasound Images

Introduction

The axillary lymph node (ALN) status of breast cancer patients is an important prognostic indicator. The use of primary breast mass features for the prediction of ALN status is rare. Two nomograms based on preoperative ultrasound (US) images of breast tumors and ALNs were developed for the prediction of ALN status.

Methods

A total of 743 breast cancer cases collected from 2016 to 2019 at the Second Affiliated Hospital of Harbin Medical University were randomly divided into a training set (n = 523) and a test set (n = 220). A primary tumor feature model (PTFM) and ALN feature model (ALNFM) were separately generated based on tumor features alone, and a combination of features was used for the prediction of ALN status. Logistic regression analysis was used to construct the nomograms. A receiver operating characteristic curve was plotted to obtain the area under the curve (AUC) to evaluate accuracy, and bias-corrected AUC values and calibration curves were obtained by bootstrap resampling for internal and external verification. Decision curve analysis was applied to assess the clinical utility of the models.

Results

The AUCs of the PTFM were 0.69 and 0.67 for the training and test sets, respectively, and the bias-corrected AUCs of the PTFM were 0.67 and 0.67, respectively. Moreover, the AUCs of the ALNFM were 0.86 and 0.84, respectively, and the bias-corrected AUCs were 0.85 and 0.81, respectively. Compared with the PTFM, the ALNFM showed significantly improved prediction accuracy (p < 0.001). Both the calibration and decision curves of the ALNFM nomogram indicated greater accuracy and clinical practicality. When the US tumor size was ≤21.5 mm, the Spe was 0.96 and 0.92 in the training and test sets, respectively. When the US tumor size was greater than 21.5 mm, the Sen was 0.85 in the training set and 0.87 in the test set. Our further research showed that when the US tumor size was larger than 35 mm, the Sen was 0.90 in the training set and 0.93 in the test set.

Conclusion

The ALNFM could effectively predict ALN status based on US images especially for different US tumor size.

Deep Learning Signature Based on Staging CT for Preoperative Prediction of Sentinel Lymph Node Metastasis in Breast Cancer

RATIONALE AND OBJECTIVES

To evaluate the noninvasive predictive performance of deep learning features based on staging CT for sentinel lymph node (SLN) metastasis of breast cancer.

MATERIALS AND METHODS

A total of 348 breast cancer patients were enrolled in this study, with their SLN metastases pathologically confirmed. All patients received contrast-enhanced CT preoperative examinations and CT images were segmented and analyzed to extract deep features. After the feature selection, deep learning signature was built with the selected key features. The performance of the deep learning signatures was assessed with respect to discrimination, calibration, and clinical usefulness in the primary cohort (184 patients from January 2016 to March 2017) and then validated in the independent validation cohort (164 patients from April 2017 to December 2018).

RESULTS

Ten deep learning features were automatically selected in the primary cohort to establish the deep learning signature of SLN metastasis. The deep learning signature shows favorable discriminative ability with an area under curve of 0.801 (95% confidence interval: 0.736-0.867) in primary cohort and 0.817 (95% confidence interval: 0.751-0.884) in validation cohort. To further distinguish the number of metastatic SLNs (1-2 or more than two metastatic SLN), another deep learning signature was constructed and also showed moderate performance (area under curve 0.770).

CONCLUSION

We developed the deep learning signatures for preoperative prediction of SLN metastasis status and numbers (1-2 or more than two metastatic SLN) in patients with breast cancer. The deep learning signature may potentially provide a noninvasive approach to assist clinicians in predicting SLN metastasis in patients with breast cancer.

Clinical value of postoperative sentinel lymph node biopsy

Background

This study analyzed the clinical data and general information of breast cancer patients who were admitted by the Affiliated Hospital of Nantong University and underwent lumpectomy, followed by sentinel lymph node biopsy (SLNB) to investigate the effect of tumor location on the sentinel lymph node (SLN) detection rate, obtain a clear understanding of the SLNB procedure and further promote the use of this procedure in the local area.

Methods

This study involved a total of 118 patients who were diagnosed with breast cancer and admitted by the Affiliated Hospital of Nantong University for lumpectomy and SLNB between July 2015 and June 2019. An analysis was conducted to explore the role of tumor location in the detection of SLNs.

Results

Tumor location was associated with the success rate of post-lumpectomy SLNB. In the case of tumor location in the upper outer quadrant (UOQ) of the breast near the axilla, the SLN detection rate was relatively low. In contrast, when a tumor occurred in any of the other quadrants or the UOQ next to the areola, the tumor location had no significant impact on the SLN detection rate. SLNB indicated that 102 out of the 118 patients had SLNs, with the detection rate of 86.4%. Particularly, for patients whose tumors were located in the UOQ near their axillae, the SLN detection rate was 30% (3/10). As to tumor location in other quadrants or the UOQ next to the areola, the SLN detection rate was up to 90.8% (99/109).

Conclusions

The performance of post-lumpectomy SLNB is associated with tumor location. SLNB is recommended when the tumor site lies in the upper inner/lower outer/lower inner quadrants (UIQ/LOQ/LIQ) of the breast or the UOQ next to the areola. If the SLNB result turns out to be negative, there is no need to perform axillary lymph node dissection (ALND). For tumor location in the UOQ of the breast, especially when it is near the axilla, SLNB is not a favorable option after lumpectomy. It is recommended that the patient receive a core needle biopsy (CNB) before SLNB.