World Journal of Surgical Oncology · 2022 · 11 citations · 26 references
MedicineOncologySurgeryBreast OncologyMetastasisSurgical OncologyBreast Cancer SurgeryBreast SurgeryLymph Node MetastasisMetastatic Breast CancerPreoperative Oncology AssessmentHigh-risk Node-negative Breast CancerPredictive Model PerformancePredictive PerformancePredictive Performance EvaluationPredictive Risk ScoringPredictive Scoring ModelsPredictive Scoring Systems
BACKGROUND: There are currently no scoring-type predictive models using only easily available pre- and intraoperative data developed for assessment of the risk of advanced axillary lymph node metastasis (ALNM) in patients with breast cancer with metastatic sentinel lymph nodes (SLNs). We aimed to develop and validate a scoring system using only pre- and intraoperative data to distinguish between non-advanced (≤ 3 lymph nodes) and advanced (> 3 lymph nodes) ALNM in patients with breast cancer with metastatic SLNs. METHODS: We retrospectively identified 804 patients with breast cancer (cT1-3cN0) who had metastatic SLNs and had undergone axillary lymph node dissection (ALND). We evaluated the risk factors for advanced ALNM using logistic regression analysis and developed and validated a scoring system for the prediction of ALNM using training (n = 501) and validation (n = 303) cohorts, respectively. The predictive performance was assessed using the receiver operating characteristic (ROC) curve, area under the curve (AUC), and calibration plots. RESULTS: Ultrasound findings of multiple suspicious lymph nodes, SLN macrometastasis, the ratio of metastatic SLNs to the total number of SLNs removed, and the number of metastatic SLNs were significant risk factors for advanced ALNM. Clinical tumor size and invasive lobular carcinoma were of borderline significance. The scoring system based on these six variables yielded high AUCs (0.90 [training] and 0.89 [validation]). The calibration plots of frequency compared to the predicted probability showed slopes of 1.00 (training) and 0.85 (validation), with goodness-of-fit for the model. When the cutoff score was set at 4, the negative predictive values (NPVs) of excluding patients with advanced ALNM were 96.8% (training) and 96.9% (validation). The AUC for predicting advanced ALNM using our scoring system was significantly higher than that predicted by a single independent predictor, such as the number of positive SLNs or the proportion of positive SLNs. Similarly, our scoring system also showed good discrimination and calibration ability when the analysis was restricted to patients with one or two SLN metastases. CONCLUSION: Our easy-to-use scoring system can exclude advanced ALNM with high NPVs. It may contribute to reducing the risk of undertreatment with adjuvant therapies in patients with metastatic SLNs, even if ALND is omitted.
26
Assessing the Performance of Prediction Models
Ewout W. Steyerberg, Andrew J. Vickers, Nancy R. Cook et al. · Epidemiology · 2009 · 4.9K citations · Full text
Antonio C. Wolff, M. Elizabeth Hammond, David G. Hicks et al. · Journal of Clinical Oncology · 2013 · 4.1K citations · Full text
Antonio C. Wolff, M. Elizabeth Hammond, Kimberly H. Allison et al. · Journal of Clinical Oncology · 2018 · 2.5K citations · Full text
Mila Donker, Geertjan van Tienhoven, Marieke E. Straver et al. · The Lancet Oncology · 2014 · 1.8K citations · Full text