A Model to Predict the Risk of Lymph Node Metastasis in Breast Cancer Based on Clinicopathological Characteristics

Assessment of the axilla in women with early-stage breast cancer undergoing primary surgery: a review

Sentinel node biopsy (SNB) is routinely performed in people with node-negative early breast cancer to assess the axilla. SNB has no proven therapeutic benefit. Nodal status information obtained from SNB helps in prognostication and can influence adjuvant systemic and locoregional treatment choices. However, the redundancy of the nodal status information is becoming increasingly apparent. The accuracy of radiological assessment of the axilla, combined with the strong influence of tumour biology on systemic and locoregional therapy requirements, has prompted many to consider alternative options for SNB. SNB contributes significantly to decreased quality of life in early breast cancer patients. Substantial improvements in workflow and cost could accrue by removing SNB from early breast cancer treatment. We review the current viewpoints and ideas for alternative options for assessing and managing a clinically negative axilla in patients with early breast cancer (EBC). Omitting SNB in selected cases or replacing SNB with a non-invasive predictive model appear to be viable options based on current literature.

Radiomic Nomogram for Predicting Axillary Lymph Node Metastasis in Patients with Breast Cancer

RATIONALE AND OBJECTIVES

The detection of axillary lymph node metastasis (ALNM) in patients with breast cancer is a crucial determinant in the decision-making process for axillary surgery and potential therapies. The objective of this study was to develop and validate a radiomics nomogram that integrates radiomics features from dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) with clinical factors to predict ALNM in patients with breast cancer.

MATERIALS AND METHODS

A total of 177 patients with breast cancer were randomly divided into a training set (n = 123) and a validation set (n = 54) using a 7:3 ratio. From the DCE-MRI images, 2818 radiomics features were extracted from the primary tumor and axillary lymph node (ALN). Subsequently, optimal features were selected through the least absolute shrinkage and selection operator algorithm to construct the Radscore. Clinical factors were identified using univariate logistic regression analysis and included in a multivariate logistic regression analysis. Using the Radscore and clinical factors, a radiomics nomogram was developed using the Support Vector Machine method. The predicting efficacy of our model was visually appraised utilizing a receiver operator characteristic (ROC) curve, while its clinical application and predictive accuracy were assessed through decision curve analysis (DCA) and calibration curves, respectively.

RESULTS

The results revealed Ki67, multifocality, and MRI-reported ALN status as independent risk factors for ALNM. The radiomics nomogram demonstrated good calibration and discrimination with areas under the ROC curve of 0.92 (95% confidence interval [CI], 0.88-0.97) in the training set and 0.90 (95% CI, 0.72-0.90) in the validation set. DCA revealed the clinical usefulness of the radiomics nomogram.

CONCLUSION

The DCE-MRI-based radiomics nomogram is a reliable tool for assessing ALNM in patients with breast cancer.

Development and validation of a point-based scoring system for predicting axillary lymph node metastasis and disease outcome in breast cancer using clinicopathological and multiparametric MRI features

BACKGROUND

Axillary lymph node (ALN) metastasis is used to select treatment strategies and define the prognosis in breast cancer (BC) patients and is typically assessed using an invasive procedure. Noninvasive, simple, and reliable tools to accurately predict ALN status are desirable. We aimed to develop and validate a point-based scoring system (PSS) for stratifying the ALN metastasis risk of BC based on clinicopathological and quantitative MRI features and to explore its prognostic significance.

METHODS

A total of 219 BC patients were evaluated. The clinicopathological and quantitative MRI features of the tumors were collected. A multivariate logistic regression analysis was used to create the PSS. The performance of the models was evaluated using receiver operating characteristic curves, and the area under the curve (AUC) of the models was calculated. Kaplan-Meier curves were used to analyze the survival outcomes.

RESULTS

Clinical features, including the American Joint Committee on Cancer (AJCC) stage, T stage, human epidermal growth factor receptor-2, estrogen receptor, and quantitative MRI features, including maximum tumor diameter, K_ep, V_e, and TTP, were identified as risk factors for ALN metastasis and were assigned scores for the PSS. The PSS achieved an AUC of 0.799 in the primary cohort and 0.713 in the validation cohort. The recurrence-free survival (RFS) and overall survival (OS) of the high-risk (> 19.5 points) groups were significantly shorter than those of the low-risk (≤ 19.5 points) groups in the PSS.

CONCLUSION

PSS could predict the ALN metastasis risk of BC. A PSS greater than 19.5 was demonstrated to be a predictor of short RFS and OS.

OSNA Total Tumor Load for the Prediction of Axillary Involvement in Breast Cancer Patients: Should We use Different Thresholds According to the Intrinsic Molecular Subtype? MOTTO Study

Aims

To assess the impact of the molecular subtype (MS) on the total number of CK19 mRNA copies in all positive SLN (TTL) threshold, to predict non-SLN affectation, and to compare 5 years progression-free survival (PFS) according to the risk of recurrence (ROR) group by PAM50.

Methods

Cohort with infiltrating breast cancer with intra-operative metastatic SLN detected by one-step nucleic acid amplification (OSNA) assay who underwent subsequent ALND. Logistic regression was used to assess a possible interaction between TTL and MS(Triple Negative, Her-2-Enriched, Luminal A, or Luminal B), or hormone receptors (HR: positive or negative) by immunohistochemistry (IMH). Cox regression was used to compare PFS and OS in the 3 ROR groups (high, medium, or low).

Results

TTL was predictive of non-SLN affectation in both univariate (OR [95% CI]: 1.72 [1.43, 2.05], P < .001) and multivariate (1.55 [95% CI: 1.04, 2.32], P = .030) models, but MS-IMH or HR-IMH, and their interactions with TTL were not (best multivariate model: HR + main effect OR 1.16 [95% CI: 0.18, 7.64], P = .874; interaction OR: 1.04 [0.7, 1.55], P = .835; univariate model: HR + main effect OR: 1.44 [95% CI: 0.85, 2.44], P = .180). PFS was lower in the high-risk ROR group (81.1%) than in the low-risk group (93.9%) (HR: 3.68 [95 CI: 1.70, 7.94], P < .001).

Conclusions

our results do not provide evidence to support the utilization of subtype-specific thresholds for TTL values to make therapeutic decisions on the axilla. The ROR group was predictive of 5 years-PFS.

A clinical diagnostic test on the detection of sentinel lymph node metastasis in breast neoplasms using a 1-step RT-PCR

BACKGROUND

Currently, assessment of sentinel lymph node (SLN) requires cytology, hematoxylin-eosin (HE), and immunohistochemistry (IHC). However, routine pathological slides still suffer from certain sampling errors and have time limitations. This study sought to investigate the sensitivity and specificity of SLN detection by reverse transcription-polymerase chain reaction (RT-PCR), which quantifies the expression of mammaglobin and cytokeratin-19 genes to determine SLN status.

METHODS

The RT-PCR detection of cycles threshold (CT) values has a direct relationship with the lymph node metastasis. This study prospectively collected 256 sentinel lymph nodes from 150 patients diagnosed with breast cancer between August and November 2017. In the detection of metastases in lymph nodes, molecular markers can be verified at the cell-level and tissue-level of tumor cells. In this study, IHC results were used as the gold standard for judging sentinel lymph node status.

RESULTS

(I) According to the established cell models, as the lymph nodes in tumor cells increase, RT-PCR CT values decrease. (II) 83 lymph nodes were first collected, and the interpretation criteria for the molecular detection results were determined based on the IHC results. (III) The statistical analysis showed that the sensitivity of the RT-PCR was 80.49% and the specificity was 91.55%. The positive predictive and negative predictive values were 64.71% and 96.06%, respectively. There was no significant difference between RT-PCR detection and IHC detection (P=0.076). Statistical chi-square analysis also showed that the difference between intraoperative freezing and immunohistochemistry was statistically significant (P=0.000). There was a statistically significant difference between intraoperative freezing and RT-PCR detection (P=0.000). RT-PCR detection is more sensitive than intraoperative frozen detection, and is closer to the results of immunohistochemistry.

CONCLUSIONS

RT-PCR had objective and rapid output advantages, and was proven to be true and reliable. RT-PCR detection can not only rapidly assess sentinel lymph node status in breast cancer patients during surgery, but its accuracy is also close to that of IHC. Correctly determine whether to perform axillary lymph node dissection and improve the survival rate of patients.

A Machine Learning-Based Predictive Model for Predicting Lymph Node Metastasis in Patients With Ewing's Sarcoma

Objective

In order to provide reference for clinicians and bring convenience to clinical work, we seeked to develop and validate a risk prediction model for lymph node metastasis (LNM) of Ewing’s sarcoma (ES) based on machine learning (ML) algorithms.

Methods

Clinicopathological data of 923 ES patients from the Surveillance, Epidemiology, and End Results (SEER) database and 51 ES patients from multi-center external validation set were retrospectively collected. We applied ML algorithms to establish a risk prediction model. Model performance was checked using 10-fold cross-validation in the training set and receiver operating characteristic (ROC) curve analysis in external validation set. After determining the best model, a web-based calculator was made to promote the clinical application.

Results

LNM was confirmed or unable to evaluate in 13.86% (135 out of 974) ES patients. In multivariate logistic regression, race, T stage, M stage and lung metastases were independent predictors for LNM in ES. Six prediction models were established using random forest (RF), naive Bayes classifier (NBC), decision tree (DT), xgboost (XGB), gradient boosting machine (GBM), logistic regression (LR). In 10-fold cross-validation, the average area under curve (AUC) ranked from 0.705 to 0.764. In ROC curve analysis, AUC ranged from 0.612 to 0.727. The performance of the RF model ranked best. Accordingly, a web-based calculator was developed (https://share.streamlit.io/liuwencai2/es_lnm/main/es_lnm.py).

Conclusion

With the help of clinicopathological data, clinicians can better identify LNM in ES patients. Risk prediction models established in this study performed well, especially the RF model.

Development of High-Resolution Dedicated PET-Based Radiomics Machine Learning Model to Predict Axillary Lymph Node Status in Early-Stage Breast Cancer

Simple Summary

Accurate clinical axillary evaluation plays an important role in the diagnosis of and treatment planning for breast cancer (BC). This study aimed to develop a machine learning model integrating dedicated breast PET and clinical characteristics for prediction of axillary lymph node status in cT1-2N0-1M0 BC non-invasively. The performance of this integrating model in identifying pN0 and pN1 with the AUC was 0.94. We achieved an NPV of 96.88% in the cN0 and PPV of 92.73% in the cN1 subgroup. The higher true positive and true negative rate could delineate clinical subtypes and apply more precise treatment for patients with early-stage BC.

Abstract

Purpose of the Report: Accurate clinical axillary evaluation plays an important role in the diagnosis and treatment planning for early-stage breast cancer (BC). This study aimed to develop a scalable, non-invasive and robust machine learning model for predicting of the pathological node status using dedicated-PET integrating the clinical characteristics in early-stage BC. Materials and Methods: A total of 420 BC patients confirmed by postoperative pathology were retrospectively analyzed. 18F-fluorodeoxyglucose (¹⁸F-FDG) Mammi-PET, ultrasound, physical examination, Lymph-PET, and clinical characteristics were analyzed. The least absolute shrinkage and selection operator (LASSO) regression analysis were used in developing prediction models. The characteristic curve (ROC) of the area under receiver-operator (AUC) and DeLong test were used to evaluate and compare the performance of the models. The clinical utility of the models was determined via decision curve analysis (DCA). Then, a nomogram was developed based on the model with the best predictive efficiency and clinical utility and was validated using the calibration plots. Results: A total of 290 patients were enrolled in this study. The AUC of the integrated model diagnosed performance was 0.94 (95% confidence interval (CI), 0.91–0.97) in the training set (n = 203) and 0.93 (95% CI, 0.88–0.99) in the validation set (n = 87) (both p < 0.05). In clinical N0 subgroup, the negative predictive value reached 96.88%, and in clinical N1 subgroup, the positive predictive value reached 92.73%. Conclusions: The use of a machine learning integrated model can greatly improve the true positive and true negative rate of identifying clinical axillary lymph node status in early-stage BC.

Identification of the lymph node metastasis-related automated breast volume scanning features for predicting axillary lymph node tumor burden of invasive breast cancer via a clinical prediction model

Breast cancer has become the malignant tumor with the highest incidence in women. Axillary lymph node dissection (ALND) is an effective method of maintaining regional control; however, it is associated with a significant risk of complications. Meanwhile, whether the patients need ALND or not is according to sentinel lymph node biopsy (SLNB). However, the false-negative results of SLNB had been reported. Automated breast volume scanning (ABVS) is a routine examination in breast cancer. A real-world cohort consisting of 245 breast cancer patients who underwent ABVS examination were enrolled, including 251 tumor lesions. The ABVS manifestations were analyzed with the SLNB results, and the ALND results for selecting the lymph node metastasis were related to ABVS features. Finally, a nomogram was used to construct a breast cancer axillary lymph node tumor burden prediction model. Breast cancer patients with a molecular subtype of luminal B type, a maximum lesion diameter of ≥5 cm, tumor invasion of the Cooper's ligament, and tumor invasion of the nipple had heavy lymph node tumor burden. Molecular classification, tumor size, and Cooper's ligament status were used to construct a clinical prediction model of axillary lymph node tumor burden. The consistency indexes (or AUC) of the training cohort and the validation cohort were 0.743 and 0.711, respectively, which was close to SLNB (0.768). The best cutoff value of the ABVS nomogram was 81.146 points. After combination with ABVS features and SLNB, the AUC of the prediction model was 0.889, and the best cutoff value was 178.965 points. The calibration curve showed that the constructed nomogram clinical prediction model and the real results were highly consistent. The clinical prediction model constructed using molecular classification, tumor size, and Cooper's ligament status can effectively predict the probability of heavy axillary lymph node tumor burden, which can be the significant supplement to the SLNB. Therefore, this model may be used for individual decision-making in the diagnosis and treatments of breast cancer.

Omitting SLNB in Breast Cancer: Is a Nomogram the Answer?

BACKGROUNDS

Sentinel lymph node biopsy (SLNB) is standard care as a staging procedure in patients with invasive breast cancer. The axillary recurrence rate, even after positive SLNB, is low. This raises serious doubts regarding the clinical value of SLNB in early breast cancer. The purpose of this study is to select patients with low suspected axillary burden in whom SLNB might be omitted.

PATIENTS AND METHODS

We retrospectively analyzed 2015 primary breast cancer patients between 2007 and 2015, with 982 patients allocated to the training and 961 to the validation cohort. Variables associated with nodal disease were analyzed and used to build a nomogram for predicting nodal disease.

RESULTS

A total of 32.8% of patients had macrometastatic disease. A predictive model was constructed based on age, cN0, morphology, grade, multifocality, and tumor size with an area under the receiver operating characteristic curve (AUC) of 0.83. Considering a false-negative rate of 5%, 32.8% of patients could be spared axillary surgery. In a subanalysis of patients with relatively favorable characteristics, 26.8% had less than 5% chance of macrometastases.

CONCLUSIONS

We present a model with excellent predictive value that can select one-third of patients in whom SLNB is deemed not necessary because of less than 5% chance of nodal involvement. Whether missing 1 in 20 patients with macrometastatic disease is worthwhile balanced against preventing side-effects of the SLN procedure remains to be established. A number of ongoing large prospective trials evaluating the outcome of omitting SLNB are awaited. Meanwhile, this nomogram may be used for individual decision-making.

Clinical Value of Three Combined Ultrasonography Modalities in Predicting the Risk of Metastasis to Axillary Lymph Nodes in Breast Invasive Ductal Carcinoma

Objective

The present study aimed to assess the clinical value of conventional ultrasound (C-US), ultrasound elastography (UE), percutaneous contrast-enhanced ultrasound (P-CUES), and the combination of these three ultrasonography modalities for evaluating the risk of axillary lymph node (ALN) metastasis in breast invasive ductal carcinoma (IDC).

Methods

This retrospective analysis included 120 patients with pathologically confirmed IDC who underwent sentinel lymph node biopsy (SLNB) or axillary lymph node dissection (ALND). Based on the gold standard of postoperative pathology, ALN pathology results were evaluated and compared with findings obtained using C-US, UE, P-CUES, and the three modalities combined.

Results

(1) There was a statistically significant difference between the histological grade of the tumor and the pathological condition of ALNs. (2) The difference between C-US parameters and UE score were statistically significant. The accuracy of P-CEUS localization of SLNs was 100% (96/96) when compared with localization guided by methylene blue. The difference in the distribution of the four SLN enhancement patterns was statistically significant. (3) The sensitivity, specificity, positive predictive value, and negative predictive value of C-US and UE were 75%, 71%, 58%, and 89%, and 71%, 72%, 50%, and 86%, respectively. The sensitivity, specificity, positive predictive value, and negative predictive value of P-CUES were 91%, 82%, 78%, 92%, respectively. When all three modalities were combined, the sensitivity, specificity, positive predictive value, and negative predictive value were 94%, 89%, 86%, and 95%, respectively. In the detection of ALN metastasis, there was a good correlation between histopathological results and evaluations based on the three combined ultrasonography modalities (kappa: 0.82, p<0.001).

Conclusions

When compared to C-US, UE, or P-CEUS alone, the combination of the three ultrasonography modalities was found to be superior in distinguishing metastatic and non-metastatic ALNs. This combined strategy may aid physicians in determining the most appropriate approach to ALN surgery as well as the prognosis of breast IDC.

Nomogram for predicting preoperative regional lymph nodes metastasis in patients with metaplastic breast cancer: a SEER population-based study

Background

Metaplastic breast cancer (MBC) is a rare subtype of breast cancer, and generally associated with poor outcomes. Lymph nodes metastasis (LNM) is confirmed as a critical independent prognostic factor and determine the optimal treatment strategies in MBC patients. We aimed to develop and validate a nomogram to predict the possibility of preoperative regional LNM in MBC patients.

Methods

MBC patients diagnosed between 1990 and 2016 in the Surveillance, Epidemiology, and End Results (SEER) database were included and stochastically divided into a training set and validation set at a ratio of 7:3. The risk variables of regional LNM in the training set were determined by univariate and multivariate logistic regression analyses. And then we integrated those risk factors to construct the nomogram. The prediction nomogram was further verified in the verification set. The discrimination, calibration and clinical utility of the nomogram were evaluated by the area under the receiver operating characteristic (ROC) curve (AUC), calibration plots and decision curve analysis (DCA), respectively.

Results

A total of 2205 female MBC patients were included in the study. Among the 2205 patients, 24.8% (546/2205) had positive regional lymph nodes. The nomogram for predicting the risk of regional LNM contained predictors of grade, estrogen receptor (ER) status and tumor size, with AUC of 0.683 (95% confidence interval (CI): 0.653–0.713) and 0.667 (95% CI: 0.621–0.712) in the training and validation sets, respectively. Calibration plots showed perfect agreement between actual and predicted regional LNM risks. At the same time, DCA of the nomogram demonstrated good clinical utilities.

Conclusions

The nomogram established in this study showed excellent prediction ability, and could be used to preoperatively estimate the regional LNM risk in MBC.