Introduction Chemo-radiotherapy (CRT) is the first line treatment in patients with anal squamous cell carcinoma (ASCC). Clinical lymph node positivity (LN+) is considered an adverse prognosticator. Accurate lymph node staging impacts on treatment, such that in LN+ patients, lymphatic fields are treated with prophylactic phase I wide volume radiotherapy. Modern imaging modalities, namely magnetic resonance imaging (MRI) and PET-CT scanning increase sensitivity to detect LN+, but It is unclear whether enhanced pre-treatment staging impacts upon oncological outcome.

Method From 1990 to 2014 we characterised the pattern of changes in the detection rates of LN+ in patients with ASCC through a dedicated anal cancer MDT serving a population of 3 million. We divided time periods as: (i) pre-MRI era; (ii) MRI era; and (iii) MRI + PET-CT era. In patients receiving curative CRT, we determined actuarial rates for loco-regional and inguinal relapses, and cancer-specific survival (CSS), using Kaplan-Meier curves. Imaging modalities were compared using log-rank tests, and Cox multivariate models that also conditioned for period and cohort effects.

Results From 971 patients with new ASSC, 550 were treated with CRT [M, F: 207:343, median age (range): 61 (29–89)]. In the pre-MRI era (pre-2004), LN+ was identified in 18% (36/200); with the use of MRI the detection rate increased to 34% (62/185); and further increased to 48% (58/120) with the use of MRI and PET-CT in combination (post-2011) (p < 0.001). In multivariate models, T stage was the only other independent determinant of LN+ (T1 2% (1/55) vs T4 55% (55/100), p < 0.001). The secular trends were associated with an altered prognostic profile. For the pre-MRI era the 5 year survival of LN+ and LN- was 51% and 78%; for the post-MRI era the 5 year survival was 75% and 85% for LN+ and LN- respectively. Additionally, there was loss of discrimination between the two prognostic categories with the significant difference in survival between the groups during the pre-MRI era being lost in the MRI-era (p = 0.0014 and 0.089 respectively) – a Will Roger phenomenon. There were minor improvements in 3-year loco-regional relapse rates over the time periods. The overall rate of inguinal nodal relapse was low (3%).

Conclusion MRI and PET scans improve pre-treatment staging in anal cancer although precise characterisation of the specific lymphatic fields involved needs further evaluation. Demonstration of the Will Roger’s effect challenges the prognostic value of pre-treatment nodal status in anal cancer and current treatment algorithms. This study was generously supported by the BDRF.

Disclosure of interest None Declared.