scholarly journals Prediction of pathologic complete response to neoadjuvant chemotherapy using machine learning models in patients with breast cancer

2021 ◽  
Author(s):  
Ji-Yeon Kim ◽  
Eunjoo Jeon ◽  
Soonhwan Kwon ◽  
Hyungsik Jung ◽  
Sunghoon Joo ◽  
...  
Keyword(s):  
Breast Cancer ◽  
Machine Learning ◽  
Neoadjuvant Chemotherapy ◽  
Pathologic Complete Response ◽  
Complete Response ◽  
Learning Models ◽  
Response To Neoadjuvant Chemotherapy ◽  
Machine Learning Models

Multi-gene prognostic assays and age-related differences in prediction of pathologic complete response to neoadjuvant chemotherapy and survival in breast cancer.

2021 ◽  
Vol 39 (15_suppl) ◽  
pp. 591-591
Author(s):  
Kent Hanson ◽  
Kent Hoskins ◽  
Naomi Yu Ko ◽  
Gregory Sampang Calip
Keyword(s):  
Breast Cancer ◽  
High Risk ◽  
Neoadjuvant Chemotherapy ◽  
Pathologic Complete Response ◽  
Complete Response ◽  
Genomic Testing ◽  
Younger Women ◽  
Her2 Breast Cancer ◽  
Response To Neoadjuvant Chemotherapy ◽  
Overall Mortality

591 Background: Multi-gene testing of primary breast tumors in early-stage breast cancer is used to classify the risk of developing distant metastases and predict the benefit of adjuvant chemotherapy. The association between the tumor genomic prognostic score (GPS) and response to neoadjuvant chemotherapy (NACT) and survival is not well characterized. Our objective was to describe the association between GPS and rates of pathologic complete response (PCR) and subsequent overall survival among women with or without PCR. Methods: We utilized the National Cancer Database to perform a hospital-based, retrospective cohort study of breast cancer patients ages 18 years and older. We included women diagnosed with first primary stages I-III hormone receptor positive (HR+), HER2 negative (HER2-) breast cancer who received NACT and surgery between 2010 and 2017. Women were categorized as having low (0-10 or 200), intermediate (11-25 or 300), or high-risk (25-199 or 400) GPS based on OncotypeDX or MammaPrint scores. Multivariable modified Poisson regression models with robust error variance were used to estimate the crude and adjusted relative risk and 95% confidence intervals (CI) for PCR associated with GPS groups. Multivariable Cox proportional hazards models were used to estimate adjusted hazard ratios (HR) and 95% CI for associations between the GPS and overall survival (OS) in women who did and did not have PCR. Results: A cohort of 3,446 women (mean [SD] age, 56.7 [12.0] years; median [interquartile range] follow-up of 47 [31-68] months) who received genomic testing and neoadjuvant chemotherapy were included in our analysis, of which 935 (27%) were low risk, 1,357 (39%) intermediate risk, and 1,154 (34%) high risk GPS. The relative risk of PCR for all women with high GPS was 1.81 (95% CI, 1.47-2.22; p < 0.001) in crude models and 1.49 (95% CI, 1.16-1.92; p = 0.002) after full adjustment compared to low GPS. Across all models, having a high GPS was significantly associated with achieving PCR in younger women ( < 65 years). In women ages ≥65 years, the association between GPS and PCR was not predictive nor statistically significantly. Among women with no response or partial response to NACT, high GPS was associated with a significantly increased risk of overall mortality (HR 2.41; 95% CI, 1.61-3.60; p < 0.001) compared to low GPS. Conversely, in women who did achieve PCR, GPS was not predictive of overall mortality across all age groups. Conclusions: In women with HR+/HER2- breast cancer, high risk GPS was predictive of PCR following NACT, primarily in younger women ( < 65 years). Our findings also indicated GPS was associated with lower OS in high-risk patients who do not achieve PCR and unpredictive of OS in those without PCR. The utility of tumor genomic testing in the neoadjuvant setting needs further investigation.

Evaluation of Tumor Stiffness by Elastography Is Predictive for Pathologic Complete Response to Neoadjuvant Chemotherapy in Patients with Breast Cancer

2012 ◽  
Vol 19 (9) ◽  
pp. 3042-3049 ◽  
Author(s):  
Mitsuhiro Hayashi ◽  
Yutaka Yamamoto ◽  
Mutsuko Ibusuki ◽  
Saori Fujiwara ◽  
Satoko Yamamoto ◽  
...  
Keyword(s):  
Breast Cancer ◽  
Neoadjuvant Chemotherapy ◽  
Pathologic Complete Response ◽  
Complete Response ◽  
Tumor Stiffness ◽  
Response To Neoadjuvant Chemotherapy

Functional proteomic-based predictor of pathologic complete response to neoadjuvant chemotherapy with paclitaxel/FAC-FEC in breast cancer

2008 ◽  
Vol 26 (15_suppl) ◽  
pp. 11078-11078
Author(s):  
A. M. Gonzalez-Angulo ◽  
B. T. Hennessy ◽  
Z. Ju ◽  
F. Meric-Bernstam ◽  
P. Lajos ◽  
...  
Keyword(s):  
Breast Cancer ◽  
Neoadjuvant Chemotherapy ◽  
Pathologic Complete Response ◽  
Complete Response ◽  
Response To Neoadjuvant Chemotherapy

Prediction of Pathologic Complete Response to Neoadjuvant Chemotherapy Using Machine Learning Models in Patients with Breast Cancer

2021 ◽  
Author(s):  
Ji-Yeon Kim ◽  
Eunjoo Jeon ◽  
Soonhwan Kwon ◽  
Hyungsik Jung ◽  
Sunghoon Joo ◽  
...  
Keyword(s):  
Breast Cancer ◽  
Machine Learning ◽  
Neoadjuvant Chemotherapy ◽  
Prediction Model ◽  
Prediction Models ◽  
Locally Advanced ◽  
Pathologic Complete Response ◽  
Complete Response ◽  
Bayesian Optimization ◽  
Pathological Characteristics

Abstract BackgroundThe aim of this study was to develop a machine learning(ML) based model to accurately predict pathologic complete response(pCR) to neoadjuvant chemotherapy(NAC) using pretreatment clinical and pathological characteristics of electronic medical record(EMR) data in breast cancer(BC).Methods The EMR data from patients diagnosed with early and locally advanced BC and who received NAC followed by curative surgery were reviewed. A total of 16 clinical and pathological characteristics was selected to develop ML model. We practiced six ML models using default settings for multivariate analysis with extracted variables. ResultsIn total, 2,065 patients were included in this analysis. Overall, 30.6% (n=632) of patients achieved pCR. Among six ML models, the LightGBM had the highest area under the curve (AUC) for pCR prediction. After hyper-parameter tuning with Bayesian optimization, AUC was 0.810. Performance of pCR prediction models in different histology-based subtypes was compared. The AUC was highest in HR+HER2- subgroup and lowest in HR-/HER2- subgroup (HR+/HER2- 0.841, HR+/HER2+ 0.716, HR-/HER2 0.753, HR-/HER2- 0.653).ConclusionsA ML based pCR prediction model using pre-treatment clinical and pathological characteristics provided useful information to predict pCR during NAC. This prediction model would help to determine treatment strategy in patients with BC planned NAC.

Sign in / Sign up

Export Citation Format

Share Document