Development of a prediction model for lymph node metastasis in luminal A subtype breast cancer: The possibility to omit sentinel lymph node biopsy

Noninvasive Staging of Lymph Node Status in Breast Cancer Using Machine Learning: External Validation and Further Model Development

BACKGROUND

Most patients diagnosed with breast cancer present with a node-negative disease. Sentinel lymph node biopsy (SLNB) is routinely used for axillary staging, leaving patients with healthy axillary lymph nodes without therapeutic effects but at risk of morbidities from the intervention. Numerous studies have developed nodal status prediction models for noninvasive axillary staging using postoperative data or imaging features that are not part of the diagnostic workup. Lymphovascular invasion (LVI) is a top-ranked predictor of nodal metastasis; however, its preoperative assessment is challenging.

OBJECTIVE

This paper aimed to externally validate a multilayer perceptron (MLP) model for noninvasive lymph node staging (NILS) in a large population-based cohort (n=18,633) and develop a new MLP in the same cohort. Data were extracted from the Swedish National Quality Register for Breast Cancer (NKBC, 2014-2017), comprising only routinely and preoperatively available documented clinicopathological variables. A secondary aim was to develop and validate an LVI MLP for imputation of missing LVI status to increase the preoperative feasibility of the original NILS model.

METHODS

Three nonoverlapping cohorts were used for model development and validation. A total of 4 MLPs for nodal status and 1 LVI MLP were developed using 11 to 12 routinely available predictors. Three nodal status models were used to account for the different availabilities of LVI status in the cohorts and external validation in NKBC. The fourth nodal status model was developed for 80% (14,906/18,663) of NKBC cases and validated in the remaining 20% (3727/18,663). Three alternatives for imputation of LVI status were compared. The discriminatory capacity was evaluated using the validation area under the receiver operating characteristics curve (AUC) in 3 of the nodal status models. The clinical feasibility of the models was evaluated using calibration and decision curve analyses.

RESULTS

External validation of the original NILS model was performed in NKBC (AUC 0.699, 95% CI 0.690-0.708) with good calibration and the potential of sparing 16% of patients with node-negative disease from SLNB. The LVI model was externally validated (AUC 0.747, 95% CI 0.694-0.799) with good calibration but did not improve the discriminatory performance of the nodal status models. A new nodal status model was developed in NKBC without information on LVI (AUC 0.709, 95% CI: 0.688-0.729), with excellent calibration in the holdout internal validation cohort, resulting in the potential omission of 24% of patients from unnecessary SLNBs.

CONCLUSIONS

The NILS model was externally validated in NKBC, where the imputation of LVI status did not improve the model's discriminatory performance. A new nodal status model demonstrated the feasibility of using register data comprising only the variables available in the preoperative setting for NILS using machine learning. Future steps include ongoing preoperative validation of the NILS model and extending the model with, for example, mammography images.

Development of prediction models for the sensitivity of oral squamous cell carcinomas to preoperative S-1 administration

S-1 is an anticancer agent that is comprised of tegafur, gimeracil, and oteracil potassium, and is widely used in various carcinomas including oral squamous cell carcinoma (OSCC). Although an established prediction tool is not available, we aimed to develop prediction models for the sensitivity of primary OSCC cases to the preoperative administration of S-1.

We performed DNA microarray analysis of 95 cases with OSCC. Using global gene expression data and the clinical data, we developed two different prediction models, namely, model 1 that comprised the complete response (CR) + the partial response (PR) versus stable disease (SD) + progressive disease (PD), and model 2 that comprised responders versus non-responders. Twelve and 18 genes were designated as feature genes (FGs) in models 1 and 2, respectively, and, of these, six genes were common to both models. The sensitivity was 96.3%, the specificity was 91.2%, and the accuracy was 92.6% for model 1, and the sensitivity was 95.6%, the specificity was 85.2%, and the accuracy was 92.6% for model 2. These models were validated using receiver operating characteristic analysis, and the areas under the curves were 0.967 and 0.949 in models 1 and 2, respectively. The data led to the development of models that can reliably predict the sensitivity of patients with OSCC to the preoperative administration of S-1. The mechanism that regulates S-1 sensitivity remains unclear; however, the prediction models developed provide hope that further functional investigations into the FGs will lead to a greater understanding of drug resistance.

Development of Prediction Model Including MicroRNA Expression for Sentinel Lymph Node Metastasis in ER-Positive and HER2-Negative Breast Cancer

BACKGROUND

The aim of our study is to find microRNAs (miRNAs) associated with sentinel lymph node metastasis (SLNM) and to develop a prediction model for SLNM in ER-positive and HER2-negative (ER+/HER2-) breast cancer.

PATIENTS AND METHODS

In the present study, only ER+/HER2- primary breast cancer was considered. The discovery set for SLNM-associated miRNAs included 10 tumors with and 10 tumors without SLNM. The training and validation sets both included 100 tumors. miRNA expression in tumors was examined comprehensively by miRNA microarray in the discovery set and by droplet digital PCR in the training and validation sets.

RESULTS

In the discovery set, miR-98, miR-22, and miR-223 were found to be significantly (P < 0.001, fold-change > 2.5) associated with SLNM. In the training set, we constructed the prediction model for SLNM using miR-98, tumor size, and lymphovascular invasion (LVI) with high accuracy (AUC, 0.877). The accuracy of this prediction model was confirmed in the validation set (AUC, 0.883), and it outperformed the conventional Memorial Sloan Kettering Cancer Center nomogram. In situ hybridization revealed the localization of miR-98 expression in tumor cells.

CONCLUSIONS

We developed a prediction model consisting of miR-98, tumor size, and LVI for SLNM with high accuracy in ER+/HER2- breast cancer. This model might help decide the indication for SLN biopsy in this subtype.

Prediction of Lymph Node Metastasis in Breast Cancer by Gene Expression and Clinicopathological Models: Development and Validation within a Population-Based Cohort

PURPOSE

More than 70% of patients with breast cancer present with node-negative disease, yet all undergo surgical axillary staging. We aimed to define predictors of nodal metastasis using clinicopathological characteristics (CLINICAL), gene expression data (GEX), and mixed features (MIXED) and to identify patients at low risk of metastasis who might be spared sentinel lymph node biopsy (SLNB).Experimental Design: Breast tumors (n = 3,023) from the population-based Sweden Cancerome Analysis Network-Breast initiative were profiled by RNA sequencing and linked to clinicopathologic characteristics. Seven machine-learning models present the discriminative ability of N0/N+ in development (n = 2,278) and independent validation cohorts (n = 745) stratified as ER⁺HER2^-, HER2⁺, and TNBC. Possible SLNB reduction rates are proposed by applying CLINICAL and MIXED predictors.

RESULTS

In the validation cohort, the MIXED predictor showed the highest area under ROC curves to assess nodal metastasis; AUC = 0.72. For the subgroups, the AUCs for MIXED, CLINICAL, and GEX predictors ranged from 0.66 to 0.72, 0.65 to 0.73, and 0.58 to 0.67, respectively. Enriched proliferation metagene and luminal B features were noticed in node-positive ER⁺HER2^- and HER2⁺ tumors, while upregulated basal-like features were observed in node-negative TNBC tumors. The SLNB reduction rates in patients with ER⁺HER2^- tumors were 6% to 7% higher for the MIXED predictor compared with the CLINICAL predictor accepting false negative rates of 5% to 10%.

CONCLUSIONS

Although CLINICAL and MIXED predictors of nodal metastasis had comparable accuracy, the MIXED predictor identified more node-negative patients. This translational approach holds promise for development of classifiers to reduce the rates of SLNB for patients at low risk of nodal involvement.

BI-RADS 3-5 microcalcifications: prediction of lymph node metastasis of breast cancer

Purpose

To determine whether the clinicopathological parameters and Breast Imaging Reporting and Data System (BI-RADS) 3–5 microcalcifications differed between lymph node positive (LN (+)) and lymph node negative (LN (−)) invasive ductal carcinoma (IDC).

Results

For microcalcification-associated breast cancers, seven selected features (age, tumor size, Ki-67 status, lymphovascular invasion, calcification range, calcification diameter and calcification density) were significantly associated with LN status (all P < 0.05). Multivariate logistic regression analysis found that three risk factors (age: older vs. younger OR: 0.973 P = 0.006, tumor size: larger vs. smaller OR: 1.671, P < 0.001 and calcification density: calcifications > 20/cm² vs. calcifications ≤ 20/cm² OR: 1.698, P < 0.001) were significant independent predictors. This model had an area under the receiver operating characteristic curve (AUC) of 0.701. The nodal staging (N0 and N1 χ² = 5.701, P = 0.017; N0 and N2 χ² = 6.614, P = 0.013) was significantly positively associated with calcification density. The luminal B subtype had the highest risk of LN metastasis. Multivariate analysis demonstrated that calcification > 2 cm in range (OR: 2.209) and larger tumor size (OR: 1.882) were independently predictive of LN metastasis in the luminal B subtype (AUC = 0.667).

Materials and Methods

Mammographic images of 419 female breast cancer patients were included. Associations between the risk factors and LN status were evaluated using a Chi-square test, ANOVA and binary logistic regression analysis.

Conclusions

This study found that age, tumor size and calcifications density can be conveniently used to facilitate the preoperative prediction of LN metastasis. The luminal B subtype has the highest risk of LN metastasis among the microcalcification-associated breast cancers.

Preoperative prediction nomogram based on primary tumor miRNAs signature and clinical-related features for axillary lymph node metastasis in early-stage invasive breast cancer

More than half patients who undergo axillary lymph node (ALN) surgery are ALN negative in early-stage invasive breast cancer (EIBC). Thus, to avoid excessive treatment, we aim to establish and validate a novel nomogram model for the preoperative diagnosis of ALN status in patients with EIBC. In total, 864 patients with EIBC from two independent centers were enrolled in our study. For the discovery set, miRNAs expression profiling with functional roles in ALN metastasis was discovered by microarray analysis and validated by quantitative polymerase chain reaction (PCR). For the training and validation cohorts, we used PCR to quantify miRNAs expression in a model development cohort and assessed miRNAs signature in an internal validation cohort and external independent validation cohort. Multivariable logistic regression analyses were used to establish a nomogram model for the likelihood of ALN metastasis from miRNAs signature and clinical variables. A signature of nine-miRNA was significantly associated with ALN status. The predictive ability of our nomogram that included miRNAs signature and clinical-related variables (age, tumor size, tumor location and axillary ultrasound-reported ALN status) was significantly greater than a model that only considered clinical-related factors (concordance index: 0.856, 0.796) and also performed well in the two validation cohorts (concordance index: 0.841, 0.747). Our nomogram is a reliable prediction method that can be conveniently used to preoperatively predict ALN status in patients with EIBC. Therefore, after further confirmation in prospective and multicenter clinical trial, omission of axillary surgery may be feasible for some patients with EIBC in the future.

Multi-gene classifiers for prediction of recurrence in breast cancer patients

Accurate prediction of recurrence risk is of vital importance for tailoring adjuvant chemotherapy for individual breast cancer patients. Although recurrence risk has been assessed by means of examination of histological data and biomarkers (ER, PR, HER2, Ki67), such conventional examinations are not accurate enough to select subsets of patients who are at sufficiently low risk of recurrence to be spared adjuvant chemotherapy without comprising the prognosis. In the past two decades or so, comprehensive gene expression analysis technology has rapidly developed and made it possible to construct recurrence prediction models for breast cancer based on multi-gene expression in tumor tissues. These models include MammaPrint, Oncotype DX, PAM50 ROR, GGI, EndoPredict, BCI, and Curebest 95GC. In clinical practice, these multi-gene classifiers are mostly used for ER-positive and node-negative breast cancer patients for whom deciding the indication of adjuvant chemotherapy based on conventional histological examination findings alone is often difficult. This article briefly reviews these multi-gene expression-based classifiers with special emphasis on Curebest™ 95GC, which was developed by us for ER-positive and node-negative breast cancer patients.