Impact of 10 Different Prior Cancer History on Survival of Patients Who Underwent Surgery for Second Primary Colorectal Cancer: a Population-based Cohort Study

In this study, the investigators collected data with the SEER*Stat software version 8.4.1 (accession number: 20377-Nov2021). The patients with confirmed diagnoses of CRC (site code C18.0-20.9 and C26.0) between 2004 and 2013 were retrieved from the Surveillance, Epidemiology, and End Results (SEER) database. Due to the significant disparity in the number of cancer survivors with various prior cancer types among newly diagnosed CRC patients, the investigators utilized the International Classification of Diseases for Oncology, 3rd edition (ICD-O-3) Site Recode to identify 10 common previous cancer sites, including the colorectum, prostate, breast, uterus, bladder, skin, lung, kidney, thyroid, and stomach. CRC patients with a prior history of cancer originated from one of the 10 sites and surgical CRC patients without a prior history of cancer were enrolled in this study. The exclusion criteria were as follows: (1) Patients with more than one cancer in the past; (2) The patient's age at diagnosis was <18 years old; (3) Patients with incomplete survival data and follow-up information; (4) Patients only had autopsy or death certificate records.

The primary outcomes of this study were overall survival (OS) and cancer-specific survival (CSS). OS was defined as the time from diagnosis to date of death（patients who were still alive at the end of follow-up were considered as censored data. CSS was defined as the time from diagnosis to date of death caused by CRC (patients who deaths from other causes or still alive at the end of follow-up were considered as censored data). The investigators set December 31, 2018, as the cut-off date for follow-up to ensure that all included cases (diagnosed in 2004-2013) were followed for at least 5 years.

Based on prior cancer history, the surgical cases were categorized into two groups: 'Non-prior cancer history' and 'Prior cancer history.' The 'Prior cancer history' group was further subcategorized by the type of prior cancer, including colorectum, prostate, breast, uterus, bladder, skin, lung, kidney, thyroid and stomach. The bias between different Prior cancer history group and Non-prior cancer history group was minimized by Propensity Score Matching (PSM), and the Kaplan-Meier method and log-rank tests were used to compare OS and CSS differences. And then, a multivariate Cox proportional hazards model was performed to estimate the hazard ratios (HR) and 95% confidence interval (CI) to analyze whether different types of prior cancer history impacted the OS and CSS in patients who underwent surgery for spCRC independently. Common demographic and clinicopathological data, including age at diagnosis of spCRC, sex, race, marital status, tumor location of spCRC, pathologic grade of spCRC, TNM stage of spCRC (0-I, II-III, IV and unknown) and chemotherapy status were entered as covariates. Kaplan-Meier curves were also constructed according to the time since first cancer diagnosis (latency), age diagnosed with spCRC, and spCRC stage.

Further analysis was performed to determine the impact of surgery on survival of spCRC patients with different type of prior cancer history. The investigators divided patients into surgical and non-surgical groups based on whether surgery was performed, and then assessed the effect of surgery on survival using the propensity score-adjusted Kaplan-Meier method.

In this study, the baseline characteristics of patients were compared using Chi-square test and Fisher's exact test. A one-to-one propensity score matching (PSM) was performed to reduce the selection bias of the two groups of baseline variables. When performing one-to-one propensity score matching between the prior cancer history group and the non-prior cancer history group, the investigators chose a caliper of 0.2. However, the investigators opted for different calipers when matching the surgical and non-surgical groups because the investigators believe that patients with different prior cancers were in entirely different situations. A two-sided probability value of P ≤ 0.05 was considered statistically significant. R software (version 4.2.3) was used for all statistical analysis.