ASO Author Reflections: Development of Natural Language Processing-Based Machine-Learning Algorithms to Identify Pathologic Complete Response from Surgical Pathology Reports

Breast cancer is the most commonly diagnosed cancer and the leading cause of cancer death among females globally.¹ Neoadjuvant (preoperative) chemotherapy (NAC) is recommended in locally advanced cases to downstage the tumor and facilitate surgical resection. Systematic reviews and meta-analyses suggest that patients who attain a pathologic complete response (pCR) after NAC achieve significantly improved overall survival.^2,3 Therefore, the presence of pCR is considered a surrogate end point for favorable long-term outcomes among breast cancer patients and plays a critical role in adjuvant systemic decision-making.⁴ Manual identification of pCR from pathology reports is extremely expensive and time-consuming. This study aimed to develop and validate a set of NLP-based algorithms that could be used to automate the detection of pCR from diagnostic biopsy pathology reports and final surgical pathology reports of breast cancer patients after NAC treatment embedded within an electronic medical record (EMR) system. …