Can AI technology augment tumor board treatment decisions for stage II colon cancer care?

Authors

null

Peng-ju Chen

Peking University Cancer Hospital and Institute, Beijing Cancer Hospital, Beijing, China

Peng-ju Chen , Ting-ting Sun , Tian-le Li , Irene Dankwa-Mullan , Alexandra Urman , Ching-Kun Wang , Yue Zhang , Yun-feng Yao , Guo-li He , Kyu Rhee , Aiwen WU

Organizations

Peking University Cancer Hospital and Institute, Beijing Cancer Hospital, Beijing, China, Qingdao Baheal Intelligent Technology Co., LTD, Qingdao, China, IBM Watson Health, Bethesda, MD, IBM Watson Health, Somers, NY, USMD Cancer Center, Fort Worth, TX, IBM Watson Health, Southbury, CT

Research Funding

Other

Background: Studies have demonstrated the benefits of adjuvant chemotherapy in patients with stage III colon cancer. However, benefits of adjuvant chemotherapy for patients with Stage II disease are less certain. Various factors, including clinicopathologic, and MSI phenotypes, are typically used to guide care, but treatment in high-risk patients remain variable. Clinical decision support (CDS) technologies armed with machine learning continue to demonstrate potential to transform standard of care, decrease treatment disparities and improve patient quality of cancer care. Methods: A cross-sectional retrospective concordance study was conducted to assess recommendations provided by IBM Watson for Oncology (WFO), a CDS system, and a multidisciplinary tumor board (MDT) decision for treatment of Stage II colon cancer. 229 patients seen at Beijing Cancer Hospital were used in the analysis. Clinical and pathological features associated with worse prognosis and defined as high risk by the MDT, included T4 primary, high-grade, poorly differentiated histology (grade 3/4 excluding MSI-H), lymphovascular invasion, bowel obstruction or perforation, perineural invasion, and inadequately sampled lymph nodes. Subgroup concordance analyses of clinicopathologic features were conducted to examine the groups that might derive some benefit from the use of WFO where evidence was limited. Results: Overall concordance was 89.1% (204/229) with high-risk subgroup results ranging from 87.5% (p = 0.68) in T4 primary to 92.7% (p = 0.02) in poorly differentiated histology. Concordance in MSI subgroup was highly significant, MSI-H 71.4% vs MSI-L 94.5% (p = 0.0). Concordance with actual treatment decisions for tumor grade was statistically significant (p = 0.02). Reasons for non-concordance included different age threshold for oxaliplatin in elderly patients and preference for capecitabine for MSI-H. Conclusions: WFO generated treatment recommendation for Stage II colon cancer can potentially augment the clinical decision-making process within MDT. Future studies are needed to assess benefits for integrating local treatment and evidence for use outside of MDT settings.

Disclaimer

This material on this page is ©2024 American Society of Clinical Oncology, all rights reserved. Licensing available upon request. For more information, please contact licensing@asco.org

Abstract Details

Meeting

2018 ASCO Annual Meeting

Session Type

Publication Only

Session Title

Publication Only: Health Services Research, Clinical Informatics, and Quality of Care

Track

Quality Care/Health Services Research

Sub Track

Clinical Informatics

Citation

J Clin Oncol 36, 2018 (suppl; abstr e18582)

DOI

10.1200/JCO.2018.36.15_suppl.e18582

Abstract #

e18582

Abstract Disclosures

Similar Abstracts

First Author: Yoanna S Pumpalova

First Author: Yoojoo Lim

Abstract

2022 ASCO Annual Meeting

High-risk MSI-H stage II colon cancer: Treatment patterns and outcomes.

First Author: Patrick Fleming