A Prospective Validation Cohort Study of a Prediction Model on Non-sentinel Lymph Node Involvement in Early Breast Cancer

• Published in: Annals of surgical oncology 2020-05, Vol.27 (5), p.1653-1658
• Main Authors: Qiao, Enqi, Yu, Xingfei, Zhou, Lingyan, Wang, Chen, Yang, Chen, Yu, Yang, Chen, Daobao, Huang, Jian, Yang, Hongjian
• Format: Article
• Language: English
• Subjects: Adult; Aged; Area Under Curve; Biopsy; Breast cancer; Breast Neoplasms - genetics; Breast Neoplasms - pathology; Breast Oncology; Cohort analysis; False Negative Reactions; Female; Humans; Keratin-19 - genetics; Lymph nodes; Lymphatic Metastasis; Lymphatic system; Medicine; Medicine & Public Health; Metastases; Middle Aged; mRNA; Multivariate Analysis; NCT; NCT02992067; NCT03280134; Nomograms; Oncology; Patients; Polymerase chain reaction; Prediction models; Predictive Value of Tests; Prospective Studies; Radiation therapy; RNA, Messenger - analysis; ROC Curve; Sentinel Lymph Node - pathology; Sentinel Lymph Node Biopsy; Surgery; Surgical Oncology; Ultrasonography; Ultrasound
• ISSN: 1068-9265; 1534-4681
• DOI: 10.1245/s10434-019-07980-x

Background Early breast cancer with one or two sentinel lymph nodes (SLNs) may omit axillary lymph node dissection (ALND) if followed by radiotherapy. However, only less than one-third of the patients have positive non-SLNs and can truly benefit from radiotherapy. Before any regional treatment decision, the risk of non-SLN metastasis must be identified. The authors previously developed a predictive model for non-SLN involvement using CK19 mRNA and contrast-enhanced ultrasound (CEUS) score in a training set. They designed a further study to evaluate the predictive effect using the model prospectively in a validation set of one or two involved SLNs. Methods This study identified early breast cancer patients at Zhejiang Cancer Hospital from July 2017 to June 2018. The CK19 mRNA tested by quantitative real-time polymerase chain reaction and CEUS scores were collected before surgery. Patients with one or two involved SLNs were enrolled and underwent ALND. The estimated percentage of non-SLN involvement was calculated by the authors’ model formula and the Memorial Sloan Kettering Cancer Center (MSKCC) nomogram. The false-negative rates, predictive accuracy, and area under curve (AUC) were compared between two predictive models. Results The study enrolled 235 patients, and 35.36% (83/235) of them had non-SLN involvement. The authors’ model had a false-negative rate of 6% and an accuracy of 94.9%. The AUC was 0.952 (95% confidence interval [CI] 0.922–0.982), which was significantly higher than that of the MSKCC model at all three cutoff value levels. Conclusion The authors’ model, using CK19 mRNA and the CEUS score, showed the potential predictive value of non-SLNs before surgery for early breast cancer patients. Clinicaltrials Registry NCT02992067, NCT03280134.