A clinical rule-based indicator to identify recurrence of colorectal cancer after curative resection using linked...

Cancer recurrence is under-recorded in most national cancer registries. We developed and validated a clinical rule-based indicator to identify recurrence after curative major resection in patients with non-metastatic colorectal cancer (CRC), based on national routinely collected administrative hospital records and chemotherapy and radiotherapy datasets.

Recurrence was defined as the cancer becoming clinically detectable again after a period of "remission" (nine months to five years after curative major resection). 34,984 CRC patients aged 18-75 years undergoing curative major resection for non-metastatic disease diagnosed between August 2014 and September 2019 in the English Cancer Registry were identified and linked to records of outpatient visits and admissions in English administrative hospital data and to chemotherapy and radiotherapy datasets. The indicator was developed with a panel of surgical and oncological experts, based on relevant diagnosis (ICD-10), procedure (OPCS-4), and administrative codes.

Of the 34,984 patients, the indicator identified 6556 (18.7 %) as having recurrence. 6173 (94.2 %) of which could be identified using administrative hospital data of admitted patients alone. Recurrence was found in a greater proportion of rectal cancer patients, and in those with more advanced T stage and N stage, and higher cancer grade. Overall and recurrence-free five-year survival from surgery was 88.7 % and 77.4 %, respectively. Two-year overall survival after recurrence was 63.9 %. 135 (82.8 %) of the 163 patients who self-reported recurrence in a national patient experience survey, and 1412 (95.2 %) of the 1483 patients with reported recurrence/progression in Cancer Registry data had recurrence defined by the developed indicator.

The validity of the CRC recurrence indicator was supported by observed associations with tumour characteristics, self-reported recurrence, and poor overall survival in patients with recurrence. This indicator can be used in research and service evaluation, to overcome the problem of incomplete cancer recurrence recording in most national cancer registries.