A meta-analysis assessing the survival implications of subclassifying T3 rectal tumours.

ABSTRACT

AIM: Although T3 tumour subclassifications have been linked to prognosis, its mandatory adoption in histopathological reports has not been incorporated. This article focusses on the survival outcomes in patients with T3 rectal cancer according to extramural spread beyond the muscularis propria. METHODS: A systematic review of all studies up to January 2016, without language restriction, was identified from MEDLINE, Cochrane Controlled Trials Register (1960-2016) and Embase (1991-2016). All studies reporting on survival and T3 tumours with a defined cut-off of 5 mm ± 1 mm tumour invasion beyond the muscularis propria for rectal cancers were included. Hazard ratios were extracted directly from the studies or from survival curves using the technique described by Parmar. Quality assessment was performed using the Newcastle-Ottawa scale. RESULTS: Tumours with invasion more than 5 ± 1 mm from the muscularis propria had statistically significantly worse overall survival (natural log of the hazard ratio [lnHR] 1.40 [1.06, 2.04], p < 0.001) and there was no statistically significant heterogeneity (χ2 = 1.541, df = 3, p = 0.673, I2 = 0). There was statistically significantly worse disease-free survival in more invasive tumours (lnHR 1.49 [1.19, 2.00], p < 0.001) and cancer specific survival (lnHR 1.22 [0.917, 1.838], p < 0.001). Overall survival in patients who had preoperative therapy was higher in patients with less invasion beyond the muscularis propria [p < 0.01]. CONCLUSIONS: Subclassifying all T3 rectal tumours according to the depth of spread with a cut-off of 5±1 mm beyond the muscularis propria is prognostically relevant for overall survival, disease-free survival and cancer-specific survival irrespective of the nodal status; therefore, subclassifying T3 tumours should be a reporting requirement in histopathology reports.