Stephanie S. Keeling, MD
T32 Research Fellow
University of Texas MD Anderson Cancer Center
Houston, Texas, United States
Chung-Yuan Hu, MPH PhD
Senior Research Scientist
University of Texas MD Anderson Cancer Center
Houston, Texas, United States
Ramy Behman, MD, PhD
Assistant Professor
University of Texas MD Anderson Cancer Center
Houston, Texas, United States
John K. Lin, MD MSHP
Assistant Professor
University of Texas, MD Anderson Cancer Center
Houston, Texas, United States
Jenny Moon, MD, MSc
Assistant Professor
University of Texas MD Anderson Cancer Center
Houston, Texas, United States
Tsuyoshi Konishi, MD, PhD
Associate Professor
University of Texas MD Anderson Cancer Center
Houston, Texas, United States
Montserrat Guraieb-Trueba, MD
Assistant Professor
University of Texas MD Anderson Cancer Center
Houston, Texas, United States
Craig A. Messick, MD
Associate Professor
University of Texas MD Anderson Cancer Center
Houston, Texas, United States
.jpg "Michael G. White, MD, MS photo")
Michael G. White, MD, MS
Assistant Professor, Director of Research
Department of Colon and Rectal Surgery, University of Texas MD Anderson Cancer Center
Houston, Texas, United States
Y. Nancy You, MD, MHSc
Professor
University of Texas MD Anderson Cancer Center
Houston, Texas, United States
J. Joshua Smith, MD, PhD
Professor and Chair
Department of Colon & Rectal Surgery, University of Texas MD Anderson Cancer Center
Houston, Texas, United States
Paula Marincola Marincola Smith, MD, PhD
Assistant Professor of Surgery
Department of Colon & Rectal Surgery, University of Texas MD Anderson Cancer Center
Houston, Texas, United States
George J. Chang, MD, MSc, MHCM
Professor
University of Texas MD Anderson Cancer Center
Houston, Texas, United States
Patients with stage II–III rectal adenocarcinoma diagnosed 2006-2022 who received neoadjuvant chemoradiation without surgery were identified from the National Cancer Database, in a retrospective design. Clinical predictors of overall survival were identified using multivariable Cox regression for risk stratification modeling. Harrell's C-index compared model discrimination based on clinically relevant predictors.
Results:
Among 22,195 patients identified, high-grade tumors (poorly differentiated: hazard ratio [HR] 1.54; 95% Confidence Interval [CI] 1.39–1.70; undifferentiated: [HR, 1.86; CI, 1.39–2.50]) and clinical T4 category [HR, 1.60; CI, 1.34–1.91] independently predicted worse survival (interaction p=0.237). Clinical N2 category, but not N1, was also associated with inferior survival [N1: HR, 1.01; CI, 0.97- 1.06; N2: HR, 1.09; CI, 1.01–1.16]. High-grade tumors were present in 18.62% of patients (poorly and undifferentiated combined), T4 category in 21.41%, and N2 category in 16.80%. A four-tier risk stratification model combining T category and grade demonstrated good discrimination (C-index 0.704). N category was excluded as it failed to improve discrimination. Compared to low-risk patients (T1-3, well/moderately differentiated), hazard ratios for death were [HR, 1.94; CI, 1.71–2.20] for T1-3 high-grade, [HR, 1.83, CI, 1.64–2.04] for T4 low-grade, and [HR, 2.78; CI, 2.36–3.28] for T4 high-grade tumors.
Conclusions:
Clinical T4 category and high tumor grade but not clinical N status were independent predictors of worse survival among patients undergoing OP for locoregional rectal cancer. Lack of prognostic discrimination by N category likely reflects poor underlying accuracy of clinical nodal assessment. Clinical T category and histologic differentiation offer meaningful prognostic value without reliance on surgical pathology and warrant consideration in future refinements of rectal cancer staging to better reflect the growing role of OP.