Cross-validation of three predictive tools for non-sentinel node metastases in breast cancer patients with micrometastases or isolated tumor cells in the sentinel node

Developing effective optimized machine learning approaches for settlement prediction of shallow foundation

The precise assessment of shallow foundation settlement on cohesionless soils is a challenging geotechnical issue, primarily due to the significant uncertainties related to the factors influencing the settlement. This study aims to create an advanced hybrid machine learning methodology for accurately estimating shallow foundations' settlement (Sm). The initial contribution of the current research is developing and validating a robust hybrid optimization methodology based on an artificial electric field and single candidate optimizer (AEFSCO). This approach is thoroughly tested using various benchmark functions. AEFSCO will also be used to optimize three useful machine learning methods: long short-term memory (LSTM), support vector regression (SVR), and multilayer perceptron neural network (MLPNN) by adjusting their hyperparameters for predicting the settlement of shallow foundations. A database consisting of 189 individual case histories, conducted through various investigations, was used for training and testing the models. The database includes five input parameters and one output. These factors encompassed both the geometric characteristics of the foundation and the properties of the sandy soil. The results demonstrate that employing effective optimization strategies to adjust the ML models' hyperparameters can significantly improve the accuracy of predicted results. The AEFSCO has increased the coefficient of determination (R2) value of the MLPNN model by 9.3 %, the SVR model by 8 %, and the LSTM model by 22 %. Also, the LSTM-AEFSCO model is more accurate than the SVR-AEFSCO and MLPNN-AEFSCO models. This is shown by the fact that R2 went from 0.9494 to 0.9290 to 0.9903, which is an increase of 4.5 % and 6 %.

Omission of Completion Axillary Lymph Node Dissection for Patients with Breast Cancer Treated by Upfront Mastectomy and Sentinel Node Isolated Tumor Cells or Micrometastases

Omission of completion axillary lymph node dissection (cALND) in patients undergoing mastectomy with sentinel node (SN) isolated tumor cells (ITC) or micrometastases is debated due to potential under-treatment, with non-sentinel node (NSN) involvement detected in 7% to 18% of patients. This study evaluated the survival impact of cALND omission in a cohort of breast cancer (BC) patients treated by mastectomy with SN ITC or micrometastases. Among 554 early BC patients (391 pN1mi, 163 ITC), the NSN involvement rate was 13.2% (49/371). With a median follow-up of 66.46 months, multivariate analysis revealed significant associations between cALND omission and overall survival (OS, HR: 2.583, p = 0.043), disease-free survival (DFS, HR: 2.538, p = 0.008), and metastasis-free survival (MFS, HR: 2.756, p = 0.014). For Her2-positive or triple-negative patients, DFS was significantly affected by cALND omission (HR: 38.451, p = 0.030). In ER-positive Her2-negative BC, DFS, OS, recurrence-free survival (RFS), and MFS were significantly associated with cALND omission (DFS HR: 2.358, p = 0.043; OS HR: 3.317; RFS HR: 2.538; MFS HR: 2.756). For 161 patients aged ≤50 years with ER-positive/Her2-negative cancer, OS and breast cancer-specific survival (BCSS) were notably impacted by cALND omission (OS HR: 103.47, p = 0.004; BCSS HR: 50.874, p = 0.035). These findings suggest a potential negative prognostic impact of cALND omission in patients with SN micrometastases or ITC. Further randomized trials are needed.

The need to tailor the omission of axillary lymph node dissection to patients with good prognosis and sentinel node micro-metastases

BACKGROUND

Results of IBCSG-23-01-trial which included breast cancer patients with involved sentinel nodes (SN) by isolated-tumor-cells or micro-metastases supported the non-inferiority of completion axillary-lymph-node-dissection (cALND) omission. However, current data are considered insufficient to avoid cALND for all patients with SN-micro-metastases.

METHODS

To investigate the impact of cALND omission on disease-free-survival (DFS) and overall survival (OS), we analyzed a cohort of 1421 patients <75 years old with SN-micro-metastases who underwent breast conservative surgery (BCS). We used inverse probability of treatment weighting (IPTW) to obtain adjusted Kaplan-Meier estimators representing the experience in the analysis cohort, based on whether all or none had been subject to cALND omission.

RESULTS

Weighted log-rank tests comparing adjusted Kaplan-Meier survival curves showed significant differences in OS (p-value = 0.002) and borderline significant differences in DFS (p-value = 0.090) between cALND omission versus cALND. Cox's regression using stabilized IPTW evidenced an average increase in the risk of death associated with cALND omission (HR = 2.77, CI95% = 1.36-5.66). Subgroup analyses suggest that the rates of recurrence and death associated with cALND omission increase substantially after a large period of time in the half sample of women less likely to miss cALND.

CONCLUSIONS

Using IPTW to estimate the causal treatment effect of cALND in a large retrospective cohort, we concluded cALND omission is associated with an increased risk of recurrence and death in women of <75 years old treated by BCS in the absence of a large consensus in favor of omitting cALND. These results are particularly contributive for patients treated by BCS where cALND omission rates increase over time.

Can a machine-learning model improve the prediction of nodal stage after a positive sentinel lymph node biopsy in breast cancer?

Background: The current standard for evaluating axillary nodal burden in clinically node negative breast cancer is sentinel lymph node biopsy (SLNB). However, the accuracy of SLNB to detect nodal stage N2-3 remains debatable. Nomograms can help the decision-making process between axillary treatment options. The aim of this study was to create a new model to predict the nodal stage N2-3 after a positive SLNB using machine learning methods that are rarely seen in nomogram development.Material and methods: Primary breast cancer patients who underwent SLNB and axillary lymph node dissection (ALND) between 2012 and 2017 formed cohorts for nomogram development (training cohort, N = 460) and for nomogram validation (validation cohort, N = 70). A machine learning method known as the gradient boosted trees model (XGBoost) was used to determine the variables associated with nodal stage N2-3 and to create a predictive model. Multivariate logistic regression analysis was used for comparison.Results: The best combination of variables associated with nodal stage N2-3 in XGBoost modeling included tumor size, histological type, multifocality, lymphovascular invasion, percentage of ER positive cells, number of positive sentinel lymph nodes (SLN) and number of positive SLNs multiplied by tumor size. Indicating discrimination, AUC values for the training cohort and the validation cohort were 0.80 (95%CI 0.71-0.89) and 0.80 (95%CI 0.65-0.92) in the XGBoost model and 0.85 (95%CI 0.77-0.93) and 0.75 (95%CI 0.58-0.89) in the logistic regression model, respectively.Conclusions: This machine learning model was able to maintain its discrimination in the validation cohort better than the logistic regression model. This indicates advantages in employing modern artificial intelligence techniques into nomogram development. The nomogram could be used to help identify nodal stage N2-3 in early breast cancer and to select appropriate treatments for patients.

Predicting the morbidity of chronic obstructive pulmonary disease based on multiple locally weighted linear regression model with K-means clustering

Chronic Obstructive Pulmonary Disease (COPD) is a common chronic respiratory disease related to inflammation affected by harmful gas and particulate matter in the air. Mathematical prediction models between COPD and air pollutants are helpful for early identification, individualized interventions to slow disease progression, and for reduction of medical expenditures. The aim was to build a regression prediction model for the occurrence of COPD acute exacerbation. We collected hospital admissions for COPD in 2015-2018 from ten hospitals in Chongqing, China, used the increment per week as response, and the local sulfur dioxide (SO2), nitrogen dioxide (NO2), carbon monoxide (CO) and particulate matter 2.5 (PM2.5) concentrations as predictor variables to build a multiple prediction model. The Mean Absolute Percentage Error (MAPE) was used to evaluate the efficiency. We found that PM2.5 and SO2 are the most important factors contributing to the improvement of prediction accuracy. Multiple locally weighted linear regression (LWLR) Model based on integrated kernel framework with the K-means algorithm demonstrated minimum prediction error of 9.03 %(k=11).

Predictive model for 5-year mortality after breast cancer surgery in Taiwan residents

BACKGROUND

Few studies of breast cancer surgery outcomes have used longitudinal data for more than 2 years. This study aimed to validate the use of the artificial neural network (ANN) model to predict the 5-year mortality of breast cancer patients after surgery and compare predictive accuracy between the ANN model, multiple logistic regression (MLR) model, and Cox regression model.

METHODS

This study compared the MLR, Cox, and ANN models based on clinical data of 3632 breast cancer patients who underwent surgery between 1996 and 2010. An estimation dataset was used to train the model, and a validation dataset was used to evaluate model performance. The sensitivity analysis was also used to assess the relative significance of input variables in the prediction model.

RESULTS

The ANN model significantly outperformed the MLR and Cox models in predicting 5-year mortality, with higher overall performance indices. The results indicated that the 5-year postoperative mortality of breast cancer patients was significantly associated with age, Charlson comorbidity index (CCI), chemotherapy, radiotherapy, hormone therapy, and breast cancer surgery volumes of hospital and surgeon (all P < 0.05). Breast cancer surgery volume of surgeon was the most influential (sensitive) variable affecting 5-year mortality, followed by breast cancer surgery volume of hospital, age, and CCI.

CONCLUSIONS

Compared with the conventional MLR and Cox models, the ANN model was more accurate in predicting 5-year mortality of breast cancer patients who underwent surgery. The mortality predictors identified in this study can also be used to educate candidates for breast cancer surgery with respect to the course of recovery and health outcomes.

Impact of completion axillary lymph node dissection in patients with breast cancer and isolated tumour cells or micrometastases in sentinel nodes

BACKGROUND

Omission of completion axillary lymph node dissection (ALND) is a standard practice in patients with breast cancer (BC) and negative sentinel nodes (SNs) but has shown insufficient evidence to be recommended in those with SN invasion.

METHODS

A retrospective analysis of a cohort of patients with BC and micrometastases (Mic) or isolated tumour cells (ITCs) in SN. Factors associated with ALND were identified, and patients with ALND were matched to patients without ALND. Overall survival (OS) and recurrence-free survival (RFS) were estimated in the overall population, in Mic and in ITC cohorts.

FINDINGS

Among 2009 patients analysed, 1390 and 619 had Mic and ITC in SN, respectively. Factors significantly associated with ALND were SN status, histological type, age, number of SN harvested and absence of adjuvant chemotherapy. After a median follow-up of 60.4 months, ALND omission was independently associated with reduced OS (hazard ratio [HR] 2.41, 90 confidence interval [CI] 1.36-4.27, p = 0.0102), but not with increased RFS (HR 1.21, 90 CI 0.74-2.0, p = 0.52) in the overall population. In matched patients, the increased risk of death in case of ALND omission was found only in the Mic cohort (HR 2.88, 90 CI 1.46-5.69), not in the ITC cohort. The risk of recurrence was also significantly increased in the subgroup of matched Mic patients (HR 1.56, 90 CI 0.90-2.73).

INTERPRETATION

A separate analysis of Mic and ITC groups, matched for the determinants of ALND, suggested that patients with Mic had increased recurrence rates and shorter OS when ALND was not performed. Our results are consistent with those of previous studies for patients with ITC but not for those with Mic. Randomised controlled clinical trials are still warranted to show with a high level of evidence if ALND can be safely omitted in patients with micrometastatic disease in SN.

Locoregional treatment of early breast cancer with isolated tumor cells or micrometastases on sentinel lymph node biopsy

The advent of sentinel lymph-node technique has led to a shift in lymph-node staging, due to the emergence of new entities namely micrometastases (pN1mi) and isolated tumor cells [pN0(i+)]. The prognostic significance of this low positivity in axillary lymph nodes is currently debated, as is, therefore its management. This article provides updates evidence-based medicine data to take into account for treatment decision-making in this setting, discussing the locoregional treatment in pN0(i+) and pN1mi patients (completion axillary dissection, axillary irradiation with or without regional nodes irradiation, or observation), according to systemic treatment, with the goal to help physicians in their daily practice.

[Regional treatment for axillary lymph node micrometastases of breast cancer]

In patients with breast cancer, axillary lymph node micrometastasis detection has been more frequent with a better definition since the introduction of the sentinel node procedure. In this review, we focus on pN1mi micrometastasis and review the literature in order to determine factors involved in making the decision of a regional treatment.

p53 expression in breast cancer predicts tumors with low probability of non-sentinel nodes infiltration

AIM

Several predictive tools of non-sentinel lymph nodes neoplastic involvement when a positive sentinel lymph node is found have been described. However, molecular factors have been rarely evaluated to build these tools. The aim of this study was to establish which factors predicted non-sentinel lymph nodes infiltration in our setting, including some molecular factors.

MATERIAL AND METHODS

We carried out a retrospective review of 161 patients with breast cancer and a positive sentinel lymph node who had undergone axillary lymph node dissection, none of whom had received neoadjuvant treatment. Features evaluated as predictive factors for non-sentinel node positivity were: menopausal status, tumor size, histological subtype, histological grade, lymphovascular invasion, extracapsular invasion, Ki67 index, hormonal receptors, CerbB2 and p53 expression, size of sentinel lymph node metastases and number of sentinel lymph nodes affected.

RESULTS

Tumor size (P = 0.001), size of sentinel lymph node metastases (P = 0.001), lobular invasive carcinoma (P = 0.05) and lymphovascular invasion (P = 0.006) were significantly associated with non-sentinel lymph node positivity. Tumor p53 positive expression was strongly associated with non-sentinel lymph node negativity (P = 0.000). In multivariate analysis, all these factors but tumor size maintained their significance. The discrimination power of the model calculated by the area under the receiver-operator curve was 0.811 (95% confidence interval, 0.741-0.880).

CONCLUSION

p53 expression in breast cancer was highly predictive of non-sentinel lymph node negativity in our study. New studies should evaluate if it would be useful to add p53 expression to other existing predictive tools.

In breast cancer patients sentinel lymph node metastasis characteristics predict further axillary involvement

The aim of the study was to correlate various primary tumor characteristics with lymph node status, to examine sentinel lymph node (SLN) metastasis size and non-SLN axillary involvement, to look for a cut-off size/number value possibly predicting additional axillary involvement with more accuracy and to examine the relationship of SLN metastasis size to overall survival. Of 301 patients who underwent SLN biopsy, 75 had positive SLNs. The size of the metastases was measured. For different size categories, association with the prevalence of non-SLN metastases was assessed. Associations between metastasis size and tumor characteristics and overall survival (OS) were studied. The prevalence of axillary lymph node (ALN) involvement was not significantly different between cases with micrometastasis or macrometastasis in SLNs (p = 0.124). However, for metastases larger than 6, 7, and 8 mm, the prevalence of ALN involvement was significantly higher (p = 0.046, 0.022, and 0.025). OS was significantly lower in SLN-positive than in SLN-negative cases (p = 0.0375). Primary tumor size larger than 20 mm was associated with a significantly higher incidence of SLN metastasis (p < 0.001), and primary tumor size over 26 mm was associated with additional positive non-SLN (p < 0.001). Higher mitotic index (≥ 7) in primary tumors was significantly (p < 0.001) associated with ALN involvement in SLN-positive cases, whereas higher Ki67 labeling index was not significantly correlated with SLN or ALN involvement. Lymphovascular invasion (LVI) in primary tumors was significantly correlated with SLN positivity (p < 0.001) but not with further ALN involvement or OS. Tumor size and LVI are predictive for SLN metastasis. Mitotic index, primary tumor size, and larger volume SLN involvement are determinants of further ALN involvement. SLN metastasis size over 6 mm is a strong predictor of further axillary involvement. OS is shorter in the presence of positive SLN.