QS223 - Characterising the risk of nodal involvement in early rectal cancer with favourable pathological features: a population-based retrospective cohort study

Image or Table (Do Not add text here):

Purpose/Background: Operative management of early rectal cancer (ERC) is influenced by the risk of nodal involvement. NCCN guidelines suggest T1 disease can be managed with local excision (LE) but recommend radical resection (RR) for T2 lesions due to a higher risk of nodal involvement (~11% in T1 vs ~24% in T2). Utilizing a population-based analysis of all patients with ERC in Ontario, Canada, we investigated the rate of nodal involvement in low risk T1 vs T2 lesions managed by RR to inform shared decision making in the management of ERC.

Methods/Interventions: All individuals with rectal adenocarcinoma diagnosed between 2010-2014 were identified from the Ontario Rectal Cancer Registry, Canada. ICD codes for diagnosis and Canadian Classification of Intervention codes for the procedure were used. Inclusion criteria were histopathologically-confirmed ERC (pT1-2N0-2, absence of perineural or lymphovascular invasion and poor differentiation) treated by RR with related surgical code (LAR, APR, Hartmann's, Proctocolectomy). Those treated with LE or neoadjuvant therapy were excluded. Data on T stage, nodal involvement and adjuvant chemotherapy were extracted in addition to age, sex, Charlson comorbidity Index (CCI) score, income quintile, residential region and surgeon volume. Multivariable logistic regression was used to explore associations between T stage, nodal involvement and demographic characteristics after adjusting for confounders established a priori. Survival (overall survival (OS)/cancer-specific survival (CSS)) were assessed using log-rank and multivariable Cox regression analysis.

Results/Outcomes: Of 1541 patients with ERC, 718 had low risk disease (T1 n=226 (32%), T2 n=492 (68%)). Most were male (65%), aged over 60 (75%) and of urban residence (78%). Nodal positivity was similar regardless of T stage (T1N+ 12.8% vs. T2N+ 17.5%, p=0.11). No differences in sex (p=0.40), CCI score (p=0.20) or surgeon volume (p=0.27) were noted between those with and without node-positive disease.

On univariable logistic regression, age (OR 0.83, 95%CI 0.78 - 0.98, p=0.02) was inversely associated with nodal positivity. T stage was not associated with N+ disease (T1 OR 0.70 (95%CI 0.44 - 1.08), T2 Ref, p=0.12). Multivariable analysis confirmed lower age to be independently associated with nodal positivity (OR 0.82 (95% CI 0.69 - 0.97), p=0.02).

Overall survival at 5 years was similar regardless of T stage: T1 83.9% (78.4%, 88.1%) vs. T2 80.7% (76.9%, 83.9%). Cancer-specific survival at 5 years was also not significantly different between groups: T1 92.2% (87.5%, 95.1%) vs T2 89.0% (85.8%, 91.6%).

Multivariable Cox regression demonstrated higher age at diagnosis (OR 2.26 (95% 1.92 - 2.65), p < 0.0001) and increasing CCI score ( < 2: Ref, 2-4: OR 1.75 (95% CI 1.21 - 2.52), >4: OR 5.82 (95% CI 2.87 - 11.79), p < 0.0001) were associated with inferior 5-year OS while female sex was associated with better 5-year OS (OR 0.71 (95% CI 0.52 - 0.97), p=0.03). T stage was not associated on univariable (T1 OR 0.92 (95% CI 0.68 - 1.24), T2 Ref, p=0.58) or multivariable (T1 OR 0.96 (95% CI 0.71 - 1.31), T2 Ref, p=0.80) analysis with OS.

Conclusion/Discussion: In this population cohort, prevalence of nodal involvement and survival outcomes were similar between low risk T1 and T2 early rectal cancer treated by radical resection. While further exploration is warranted, these data suggest that differentiating treatment of low risk ERC based on the risk of nodal involvement may not be justified.