Download This Paper
Preprints with The Lancet is a collaboration between The Lancet Group of journals and SSRN to facilitate the open sharing of preprints for early engagement, community comment, and collaboration. Preprints available here are not Lancet publications or necessarily under review with a Lancet journal. These preprints are early-stage research papers that have not been peer-reviewed. The usual SSRN checks and a Lancet-specific check for appropriateness and transparency have been applied. The findings should not be used for clinical or public health decision-making or presented without highlighting these facts. For more information, please see the FAQs.
Machine Learning for Predicting Response to Neoadjuvant Chemotherapy in Triple-Negative Breast Cancer: A Systematic Review and Meta-Analysis
42 Pages Posted: 27 Sep 2023
More...Abstract
Background: Using medical images, machine learning (ML) techniques can provide distinct advantages in predicting the efficacy of neoadjuvant chemotherapy (NACT) for breast cancer. However, there is limited evidence regarding their performance in patients with triple-negative breast cancer (TNBC). To address this gap, we conducted a comprehensive systematic review and meta-analysis to assess the predictive capabilities of ML models utilizing medical images for estimating pathological complete response (pCR) after NACT in TNBC patients.
Methods: A systematic search of Embase, PubMed, and Web of Science databases was conducted to identify pertinent studies published until 12 April 2023. The inclusion criteria were studies that developed or utilized ML techniques for predicting pCR from medical images of TNBC patients. The primary outcome was the performance of the models by pooled Area Under the Curve (AUC) using a random effects model. Moreover, we pooled the AUC in the tuning datasets and validation test datasets, respectively. The Prediction Model Risk of Bias Assessment Tool was selected to assess the potential risk of bias. The protocol was registered on PROSPERO (CRD42023442615).
Findings: Our meta-analysis included six articles with a total of 1334 patients, with a pooled AUC of 0·71 [95% confidence interval: 0·65-0·78]. Subgroup analysis revealed a superior predictive performance of the model based on MRI images compared to whole slide imaging (AUC: 0·84 [0·81-0·86] vs. 0·62 [0·59-0·65]). Furthermore, the ML models demonstrated higher predictive performance in the tuning datasets when compared to the validation test datasets (AUC: 0·81 [0·73-0·89] vs.0·69 [0·64-0·74]).
Interpretation: Image-based ML models show significant potential for predicting pCR after NACT in TNBC patients. Notably, utilizing MRI images may enhance prediction accuracy. Nevertheless, to facilitate clinical integration and improve therapeutic decision-making in TNBC patients, further refinement and validation of our findings through larger prospective studies are essential.
Funding: The Key-Area Research and Development Program of Guangdong Province, China (No.2021B0101420006); Guangdong Provincial Key Laboratory of Artificial Intelligence in Medical Image Analysis and Application (No.2022B1212010011); National Natural Science Foundation of China (No.82272088, 82071892, 82271941); The Natural Science Foundation for Distinguished Young Scholars of Guangdong Province (No.2023B1515020043).
Declaration of Interest: The authors declare no competing interests.
Keywords: Triple-negative breast cancer, Neoadjuvant chemotherapy, Machine learning, Medical image, Pathological complete response, Meta-analysis
Suggested Citation: Suggested Citation