ObjectiveTo implement artificial neural network (ANN) algorithms for noninvasive lymph node staging (NILS) to a decision support tool and facilitate the option to omit surgical axillary staging in breast cancer patients with low-risk of nodal metastasis.MethodsThe NILS tool is a further development of an ANN prototype for the prediction of nodal status. Training and internal validation of the original algorithm included 15 clinical and tumor-related variables from a consecutive cohort of 800 breast cancer cases. The updated NILS tool included 10 top-ranked input variables from the original prototype. A workflow with four ANN pathways was additionally developed to allow different combinations of missing preoperative input values. Predictive performances were assessed by area under the receiver operating characteristics curves (AUC) and sensitivity/specificity values at defined cut-points. Clinical utility was presented by estimating possible sentinel lymph node biopsy (SLNB) reduction rates. The principles of user-centered design were applied to develop an interactive web-interface to predict the patient’s probability of healthy lymph nodes. A technical validation of the interface was performed using data from 100 test patients selected to cover all combinations of missing histopathological input values.ResultsANN algorithms for the prediction of nodal status have been implemented into the web-based NILS tool for personalized, noninvasive nodal staging in breast cancer. The estimated probability of healthy lymph nodes using the interface showed a complete concordance with estimations from the reference algorithm except in two cases that had been wrongly included (ineligible for the technical validation). NILS predictive performance to distinguish node-negative from node-positive disease, also with missing values, displayed AUC ranged from 0.718 (95% CI, 0.687-0.748) to 0.735 (95% CI, 0.704-0.764), with good calibration. Sensitivity 90% and specificity 34% were demonstrated. The potential to abstain from axillary surgery was observed in 26% of patients using the NILS tool, acknowledging a false negative rate of 10%, which is clinically accepted for the standard SLNB technique.ConclusionsThe implementation of NILS into a web-interface are expected to provide the health care with decision support and facilitate preoperative identification of patients who could be good candidates to avoid unnecessary surgical axillary staging.

Background:Most patients diagnosed with breast cancer present with node–negative disease. Sentinel lymph node biopsy (SLNB) is routinely used to stage the axilla, leaving patients with healthy axillary lymph nodes without therapeutic effects but at risk of morbidities from the intervention. Numerous studies have developed nodal (N) 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 N metastasis; however, it is challenging to assess preoperatively.Objective:To externally validate a multilayer perceptron (MLP) model for noninvasive lymph node staging (NILS) in a large population-based register cohort (n=18 633) while developing a new N 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. Furthermore, we aimed to develop and validate an LVI MLP to predict missing values of LVI status to increase the preoperative feasibility of the original NILS model.Methods:Three non-overlapping cohorts were used for model development and validation. Four N MLPs and one LVI MLP were developed using 11–12 routinely available predictors. Three N models were used to account for the different availabilities of LVI status in the cohorts and external validation in NKBC. The fourth N model was developed for 80% of NKBC cases (n=14 906) and validated in the remaining 20% (n=3727). Three alternatives for imputing missing values of the LVI status were compared using the LVI model. The discriminatory capacity was evaluated by validation area under the receiver operating characteristics curve (AUC) in three of the N 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 N models. The new N model was developed in NKBC without information on LVI (AUC 0.709, 95% CI: 0.688–0.729) with excellent calibration in the hold-out internal validation cohort, resulting in the potential omission of 24% of patients from unnecessary SLNB.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 N 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.

While breast cancer commonly presents at an early stage without axillary lymphatic spread, surgical axillary staging by sentinel lymph node biopsy (SLNB) is routinely performed in all patients with clinically node-negative disease. This study presents the development and implementation of a personalized prediction tool for noninvasive lymph node staging (NILS) using artificial neural network algorithms. Routinely available clinical and tumor-related variables in the preoperative setting from a consecutive cohort of 800 breast cancer cases were included in model training and internal validation. The model was trained to handle missing data for the ten input variables, and the performance to distinguish benign/metastatic lymph nodes ranged from AUCs 0.72 to 0.74, with good calibration. The potential to abstain from axillary surgery was observed in 26% of patients using NILS for the prediction of nodal status and acknowledged a false negative rate of 10%, which is clinically accepted for the standard SLNB technique.

Background: Lymphovascular invasion (LVI) is one of the most important predictors for nodal status in breast cancer patients [1]. Multiple models have been published for prediction of preoperatively disease-free axillary using i.a. LVI [1-2]. However, LVI detection in preoperative core needle biopsy has been reported with a failure rate of 30% [3] and the analysis is not routinely performed in Sweden. Thus, a preoperative model of LVI status would be useful in prediction models for noninvasive lymph node staging (NILS). The purpose of this study was to develop an artificial neural network (ANN) model for LVI prediction using only clinicopathological variables that are routinely available in the preoperative setting. Methods: Data gathered prospectively during 2009-2012 in Lund, Sweden from 761 clinically node negative breast cancer patients were retrospectively extracted. Inclusion criteria were female sex, primary breast cancer and that each patient was scheduled for primary surgery. Patients with metastatic disease, bilateral cancer, tumor size greater than 50 mm, previous ipsilateral breast or axillary surgery, patients omitted of standard axillary staging procedure by SLNB or ALND, and those who had neoadjuvant treatment were excluded. LVI was assessed on surgical breast specimens and was defined as the presence of tumor cells within endothelium-lined vascular channels. Out of the 761 patients in the cohort, 613 patients were documented with LVI status. The LVI full case dataset was split 80/20 for training and validation. The remaining 148 patients were set aside for model testing. Since the test dataset did not contain information on LVI status, it was used to compare the predicted fraction of LVI positive patients to that of the development dataset. Only variables possible to obtain in the preoperative setting were included in the prediction models, comprising age, menopausal status, mode of detection (mammography screening or symptomatic representation), tumor size, multifocality (yes/no), histopathological type, histological grade, ki-67 percentage, estrogen receptor status, progesterone receptor status, and human epidermal growth factor receptor 2 status. An ensemble approach was used, where each ensemble constituted 30 ANNs that were trained and validated using 5-fold cross validation. For every ensemble model, different model parameters, such as L2-regularization and the number of hidden nodes, were tested. Model selection was based on validation AUC. Results: The study cohort included female clinically node negative breast cancer patients scheduled for primary surgery. Data from 613 patients (lymph node stages N0: 67.4%, N1: 26.9%, N2+: 5.7%) were used to develop the model, and 148 patients (N0: 56.8%, N1: 35.8%, N2+: 7.4%) constituted the internal test cohort. Fifteen percentage of the patients in the development dataset were LVI positive. The selected ensemble model achieved a validation AUC of 0.80 (CI 0.75-0.85). This model predicted an LVI positive rate of 16.2% in the test dataset. Conclusion: LVI was predicted with high accuracy using an ANN model based on routine preoperative clinicopathological variables. The result of validation AUC 0.80 (CI 0.75-0.85) indicates a potential for preoperative prediction of LVI, and the model can putatively be useful when applying preoperative nodal prediction models in patients without known LVI status. To confirm these results, verification in an external dataset is needed. Validation of the LVI-model in an independent dataset from the National Breast Cancer Registry will be performed, as well as an evaluation of the usefulness of the LVI-model as an imputation in a nodal prediction model. [1] Dihge, L. et al. BMC Cancer (2019). PMID: 31226956 [2] Bevilacqua, J. L. et al. J Clin Oncol. (2007). PMID: 17664461 [3] Harris, G. C. et al. Am J Surg Pathol. (2003). PMID: 12502923 Citation Format: Malin Hjärtström, Looket Dihge, Pär-Ola Bendahl, Mattias Ohlsson, Lisa Rydén. Steps toward noninvasive lymph node staging (NILS) in clinically node negative patients: Artificial neural network model to preoperatively predict lymphovascular invasion [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr P5-01-08.