Between July 1996 and March 2019, there were 20,714 patients who were diagnosed with BrC in the Severance Hospital and Gangnam Severance Hospital, Seoul, Korea. Among them, 166 patients who were also diagnosed with CRC were evaluated retrospectively. Patients who did not undergo curative resection were histologically not confirmed or had metastases at the time of diagnosis. Patients who were diagnosed with both breast and CRC synchronously within 6 months were defined as belonging to the B=C group. BrC patients who were diagnosed with CRC more than 6 months later were defined as the B-first group, while the CRC patients who were diagnosed with BrC at least 6 months later were defined as the C-first group (Fig. 1). A total of 105 patients were divided into the B=C group of 21 patients and metachronous groups of 84 patients according to the time of diagnosis of each cancer. Among the metachronous double primary cancer patients, there were 40 patients in the B-first group and 44 patients in the C-first group (Fig. 2). Data were gathered from the electronic medical recording system of Yonsei University Health System and were analyzed retrospectively. The study was carried out in accordance with the Declaration of Helsinki and was approved by the Institutional Review Board of Gangnam Severance Hospital (No. 3-2021-0165) with a waiver for informed consent.

The data of patient characteristics were collected such as sex, age at diagnosis, body mass index (BMI) at diagnosis, comorbidities, past medical histories, and family histories. Clinicopathologic variables related to BrC such as tumor location, type of histology, T stage, lymph nodes (LN) metastasis, type of adjuvant treatments, TNM stage, estrogen receptor (ER), progesterone receptor (PR), Ki-67, human epidermal growth factor receptor 2 (HER2), and type of molecular classification were especially noted. In the case of CRC, tumor location, type of surgical treatment, histologic differentiation, T stage, LN metastasis, and TNM stage were evaluated. Molecular classification of BrC was performed using the surrogate definitions from St. Gallen Consensus 2013 [15]. TNM stage of each cancer was determined according to American Joint Committee on Cancer cancer stage manual, 8th edition [16].

The associations of the 3 different groups and the continuous variables were evaluated using analysis of variance or independent 2-sample t-test. The chi-square or Fisher exact test was used to compare clinicopathological characteristics of patients for dichotomous variables. Overall survival was calculated from the date of diagnosis of the first cancer. Survival curves were plotted using the Kaplan-Meier method, and group differences in survival were evaluated by the log-rank test. Cox proportional hazard model was used to identify variables that were independently associated with overall survival. All statistical tests were 2-sided, and P<0.05 was considered statistically significant. Statistical analyses were carried out using the IBM SPSS Statistics ver. 25.0 (IBM Corp., Armonk, NY, USA).

In the C-first group, there was only 1 male patient. The mean ages at diagnosis of the first primary cancer were 63.0 years in patients of the B=C group, 57.0 years for the B-first group, and 59.6 years for the C-first group. The ages upon diagnosis of the second primary cancer were 62.5 years in the B-first group and 64.7 years in the C-first group. There were no significant differences among the 3 groups regarding the age upon diagnosis of both the first and second primary cancers. The intervals of diagnosis between breast and CRC were 67.1 months in the B-first group and 59.3 months in C-first group. While the patients of the B-first group had higher rates of diabetes mellitus, there was no significant difference for BMI, previous cancer histories, and familial cancer histories among the groups. Carcinoembryonic antigen (CEA) level at the diagnosis of BrC in the B=C group was 8.1 ng/mL and 8.5 ng/mL for CRC. Tumor markers for CEA and carcinoma antigen 15-3 at the diagnosis of BrC and CRC were significantly different among the groups (Table 1).

The tumor location of BrC was not significantly different among the groups. There was 1 patient each with bilateral BrC in both the B-first group and the C-first group. Invasive ductal carcinoma was the most common histologic type in all groups. T1 stage for BrC was 52.4 %, 45.0%, and 61.4% for the B = C group, B-first group, and C-first group, respectively. The rate of LN metastasis was 14.3%, 32.5%, and 20.5% for the B=C group, B-first group, and C-first group, respectively. For the TNM stage, stage I was most the common, followed by stage II for BrC. In the B-first group, 57.5% received chemotherapy, which was higher than the B=C or C-first group (P<0.001). The expression of ER, PR, and HER2 and the median value of Ki-67 were not significantly different among the 3 groups. For the molecular classification, the rate of patients with triple negative was 60.0% in the C-first group, 30% in the B-first group, and 9.5% in the B=C group. However, there was no significant difference for molecular classification among groups (Table 2).

Regarding tumor location of CRC, a left-sided tumor involving rectal cancer was more common than right-sided cancer. The rate of rectal cancer was 23.8%, 37.5%, and 40.9% for the B=C group, B-first group, and C-first group, respectively. In the B-first group, 20% of the patients were treated by endoscopic submucosal resection, while for the B=C group and C-first group, 95.2% and 95.5% received surgical treatments, respectively (P=0.044). Adenocarcinoma was the most common histologic type in all groups. T3 was the most common T stage, with no significant difference among the groups. The rate of regional LN metastasis was 42.9% in the B=C group, 22.5% in the B-first group, and 29.5% in the C-first group. However, there were no significant differences in tumor location, treatment, histologic type, and TNM stages for CRC (Table 3).

The median follow-up period was 13.9 years. The survival outcomes of B=C, B-first, and C-first groups are shown in Fig. 3A. In the survival analysis for all groups, the survival of the B-first group was poorer than the other groups with marginal significance (P=0.054). In the post hoc analysis, the overall survival between the B-first and C-first groups was significantly different (P=0.039). In contrast, the B = C group showed no significant difference in the survival outcomes with the B-first and C-first groups. In the survival comparisons between synchronous and metachronous cancers for BrC and CRC, there were no statistically significant differences between the 2 groups (P=0.319). In comparing the metachronous cancers between the B-first and C-first groups, the survival of the C-first group was better than the B-first group as shown in Fig. 3C (P=0.039). According to the TNM stages of CRC, there was no significant difference in survival among the 3 groups: stage I (P=0.407), stage II (P=0.346), and stage III (P=0.565) (Supplementary Fig. 1).

In the univariate analysis for overall survival, there were no significant differences among the groups (hazard ratio [HR] of C-first group, 0.23 and B=C group, 0.22; P=0.086 and P=0.364, respectively). Ages at cancer diagnosis, tumor markers, histologic types, and TNM stages in both BrC and CRC showed no significant relations to overall survival. On the other hand, HER2 positivity increased the HR by 11.9 times for the overall survival (P=0.032). In addition, regional LN metastasis of CRC increased the HR by 5.81 times (P=0.036). However, hormone receptors and the molecular classification of BrC showed no significant outcomes affecting the overall survival. In the multivariate analysis, the HR of HER2 positivity was 7.07 and 5.44 for the CRC LN metastasis. However, the values for these variables were not found to be statistically significant (Table 4).