Introduction: Although clinicopathologic staging of colorectal cancer clearly separates cancers according to prognosis, up to 20% of early stage rectal cancers recur after “curative” surgery. These patients usually do not receive adjuvant therapy because of their supposedly good prognosis. Attempts at improving the accuracy of clinicopathologic staging have generally been ineffective. This study attempts to refine prognosis of early stage rectal cancers by constructing a preoperative prognostic nomogram.
Methods: Clinical data and oncologic follow-up for 571 patients with Stage I and II rectal cancer treated by surgical resection at a single institution were modeled for prediction of disease recurrence. Patients receiving preoperative radiation or chemotherapy were excluded. Treatment failure was recorded as clinical evidence of disease recurrence or initiation of adjuvant chemotherapy or radiotherapy. Clinical data included patient age, gender, symptoms, family history of colorectal cancer, CEA, ASA class, hemoglobin, tumor size, T stage, and histologic tumor differentiation. Predictors were used in Cox proportional hazards regression which formed the basis of a nomogram to predict the probability of disease recurrence. Accuracy of the nomogram was assessed by discrimination and calibration.
Results: The 5-year probability of treatment failure for the cohort was 15%. Disease recurrence was noted in 57 of 571 patients. Eighteen recurrences were local only, 39 recurrences were distant only, and 3 were both local and distant. An additional 29 patients received adjuvant therapy postoperatively. Patients without failure had a median follow-up of 20 months (range, 0-118 months). The predictions from the nomogram appeared accurate and discriminating, with a bootstrap corrected concordance index of 0.679.
Conclusions: A preoperative nomogram has been developed and can be used to predict the probability of treatment failure among patients with early stage node-negative rectal cancer. This information can be used in discussions with patients about prognosis and may be considered in decisions regarding neoadjuvant therapy.