Abstract PS6-28: Development and external validation of a clinical nomogram for individually predicting survival of metaplastic breast carcinoma

Abstract

Abstract Background: Metaplastic breast cancer (MBC) is a rare subtype with distinct clinicopathological features, but few studies have concerned the prognosis of this rare and morphologically diverse malignancy. A prognostic index estimating the clinical outcomes for MBC would be attractive in current clinical practice. The aim of the present study was to develop and validate an effective nomogram for predicting survival of MBC. Methods: In current study, we retrospectively analyzed 1017 MBC patients between January 2005 and December 2015 from the Surveillance, Epidemiology, and End Results (SEER) database. Prognostic factors were identified based on Cox regression analyses, and the final clinical nomogram was developed to predict the 1-, 3-, or 5-year overall survival (OS). The model was validated using bootstrap resampling of SEER validation set and a Chinese cohort for external validation. Calibration curves were provided to internally validate the performance of the nomogram and discriminative ability was appraised by concordance index (C-index). Results: A total of 1017 MBC patients were included for our analysis,whose clinical and tumor characteristics were shown in Table 1. Patients diagnosed between 2005 and 2015 were assigned into 3:1 as training set (n=763) and SEER validation set (n=254). An external validation was performed by an individual set of 94 MBC patients from National Cancer Center in China from 2005 to 2018. The nomogram finally consisted of seven independent prognostic factors including age, tumor grade, tumor size, AJCC Node stage, surgery, chemotherapy and radiotherapy. The nomogram presented a good accuracy for predicting the OS with the C-index of 0.77. Interestingly, the nomogram based on the western (including 92.5 %non-Asian) SEER validation population(C-index of nomogram: 0.76)also had an optimal discrimination in Asian population (C-index of nomogram: 0.70).The calibration plots of the nomogram predictions was also accurate and corresponded closely with the actual survival rates. Conclusions: In conclusion, this novel nomogram was accurate enough to predict the OS by using readily available clinicopathologic factors in MBC general population. This prognostic index proposed here could provide individualized recommendations for patients and clinical decisions for physicians. Table 1. Characteristic of primary metaplastic breast carcinoma cohort from SEER databaseCharacteristicsPatient with MBC (N=1017)%Univariable Analysis PAge, years59.9±13.20.001<6049448.6≥6052351.4Race0.331White71970.7Black22221.8Others767.5Laterality0.374Left51850.9Right49949.1Tumor grade0.001Grade I, II15014.7Grade III, IV86785.3Primary tumor size0.001<5cm80278.9≥5cm21521.1Pathological type0.753Metaplastic carcinoma91590.0Squamous cell carcinoma484.7Spindle cell carcinoma252.5Malignant myoepithelioma70.7Adenocarcinoma mixed222.1Molecular Subtype0.696Non-TNBC31130.6TNBC70669.4AJCC N stage0.001N075774.4N117417.1N2,N3868.5Primary tumor resection0.001No525.1Yes96594.9Radiotherapy0.001No53252.3Yes48547.7Chemotherapy0.003No28628.1Yes73171.9 Citation Format: Jiani Wang, Qing Li, Yang Luo, Fei Ma, Ruigang Cai, Qiao Li, Ying Fang, Jiayu Wang, Peng Yuan, Pin Zhang, Binghe Xu. Development and external validation of a clinical nomogram for individually predicting survival of metaplastic breast carcinoma [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr PS6-28.