EP558 - Concordance Between Clinical and Pathological Staging in Colon Adenocarcinoma: Are We Overconfident in Our Clinical Assessments?

Image or Table (Do Not add text here):

Purpose/Background: Complete mesocolic excision (CME) has been proposed to improve oncological outcomes in select patients with colon cancer, yet there is ongoing debate regarding its implementation and the necessity of central vascular ligation (CVL). Accurate prediction of tumor stage and the presence of nodal metastases are essential for effective pre-therapeutic locoregional staging. However, discrepancies between clinical assessments and pathological findings raise concerns about the reliability of preoperative staging. This study aims to critically evaluate the extent of correlation between preoperatively reported clinical TN staging and pathological TN staging in a representative U.S. population, focusing on identifying factors that contribute to these disparities and questioning the confidence we place in our clinical assessments.

Methods/Interventions: This retrospective study utilized data from the National Cancer Database (NCDB), spanning 2018 to 2020, focusing on the AJCC 8th edition. The study population comprised patients diagnosed with colon adenocarcinoma who underwent surgery or polypectomy. Correlations between clinical T and N stages and their corresponding pathologic stages were evaluated using kappa statistics. Specifically, the analysis assessed factors influencing the discrepancy between up-staged and no-change groups in the cN0 classification (comparing clinical N0 to pathologic N1 to N2b) and in the cT1pN0 classification (comparing clinical T1 to pathologic T2 to T4b).

Results/Outcomes: The study included 11,151 patients with known cT and pT stages and 47,237 with known cN and pN stages. Clinical T and N stages were concordant with pathologic stages in 70.5% and 90.0% of patients, respectively. Agreement between clinical and pathologic stages was moderate for T stage (kappa = 0.572, p < 0.001) and slight for N stage (unweighted kappa = 0.183, p < 0.001). However, when weighted for cN0 and pN0 agreement, N stage showed almost perfect agreement (weighted kappa = 0.887, p < 0.001). Clinical under-estimation of nodal status occurs in 7.7% of cases, while over-estimation occurs in 2%, potentially leading to under- or overtreatment.
Independent predictors of N stage discrepancy included age >65 years (OR: 1.97, P < 0.001), Charlson-Deyo Score >1 (OR: 1.18, P < 0.001), overlapping colon lesions (OR: 1.78, P = 0.02), pT4 stage (OR: 37.71, P < 0.001), poorly differentiated adenocarcinoma (OR: 6.20, P < 0.001), lymphovascular invasion (OR: 9.95, P < 0.001), and perineural invasion (OR: 5.96, P < 0.001). For cT1pN0 classification discrepancies, distinct colon cancer primary locations and mucinous carcinoma histology (OR: 3.4, P < 0.001) were identified as independent predictors when comparing clinical T1 to pathologic T2-T4b.

Conclusion/Discussion: In a representative U.S. population, agreement between clinical and pathologic stages was moderate for T stage and slight for N stage in colon cancer patients. Various patient and tumor characteristics significantly influence the accuracy of clinical staging. Enhancing the precision of preoperative staging is crucial for optimizing surgical decision-making and improving patient outcomes.