Shorter survival in adolescent and young adult patients, compared to adult patients, with stage IV colorectal cancer in Japan

Subjects were 861 consecutive patients who were diagnosed with stage IV CRC at the age of 15 to 74 years, and who were referred to the surgery or gastrointestinal oncology division of the National Cancer Center Hospital between January 1997 and December 2013. Patients who were initially diagnosed to have stage IV CRC with histologic diagnoses of adenocarcinoma were included. Other histological types were excluded. Since our analysis focused on a comparison of adolescent and young adult patients (aged 15–39 years) [4] with adult patients (aged 40–74 years), we excluded patients aged ≥75 or < 15 years. The reason why excluded the aged patients (≥75 years) is that, because the evidence of the safety and effectiveness of oxaliplatin-based chemotherapy is lacking in the Asian population, systemic chemotherapy without oxaliplatin are recognized as standard treatments in Japan, which is clearly different from that for young and adult patients.

This retrospective study was approved by the Institutional Review Board (IRB) of the National Cancer Center Hospital (IRB code: 2015–320).

Among stage IV CRC, prognosis of patients who underwent curative resection differed from that of patients with unresectable stage IV CRC. Moreover, among unresectable stage IV CRC, palliative resection of the primary tumor may be associated with improved OS as we reported previously. [15] Thus, the subjects were classified into the following subgroups: patients who underwent primary tumor resection with metastasectomy (i.e., curative resection), patients who underwent primary tumor resection without metastasectomy (i.e., palliative resection of the primary tumor), and patients who did not undergo resection. Curative resection included R0 resection for peritoneal metastasis as we reported previously. [16]

The following parameters were assessed in the medical records retrospectively: age, treatment year, gender, Eastern cooperative oncology group performance status (PS), symptoms (asymptomatic/symptomatic), location of primary tumor (proximal colon including the cecum, hepatic flexure, and transverse colon; distal colon including the splenic flexure, sigmoid, and rectosigmoid junction; and rectum), histological type, and serum carcinoembryonic antigen (CEA) levels before treatment. The M category was assessed according to the Union for International Cancer Control TNM classification (8th edition), which was recently revised to include the following three subcategories: M1a (metastasis to one organ, excluding peritoneum), M1b (metastasis to more than one organ, excluding peritoneum), and M1c (metastasis to the peritoneum, with or without other organ involvement). [17] The grade of liver metastasis was also assessed according to the Japanese classification of colorectal carcinoma by the Japan Society for Cancer of the Colon and Rectum [18]. In this Japanese classification, grade of liver metastasis is determined according to the number and the maximum diameter of the metastatic lesion(s), and H3 is defined as diffuse liver metastases (≥ 5 lesions) with the maximum diameter of hepatic metastases > 5 cm.

Follow-up consisted of serum tumor marker measurements every one to three months and computed tomography (CT) every three to six months, with cut-off date of July 2017. Complete follow-up was conducted for the entire cohort of patients, with a median follow-up time for survivors of 47 months (range, 2–144 months).

Pearson’s chi-square test for categorical variables and the Wilcoxon rank-sum test for continuous variables were used to compare various factors in both groups. OS was defined as the interval between the date of stage IV CRC diagnosis and the date of death for all-cause, and censored for survivors at the date of the data cut-off. The Kaplan-Meier method was used to estimate OS. Differences in survival outcomes were assessed by the log-rank test. Multivariate Cox proportional hazards regression models were performed to evaluate the prognostic impact of age (adolescent and young adult or adult) with OS, adjusting for key clinicopathological factors (treatment year, PS, symptoms, primary tumor location, M category, grade of liver metastasis, tumor differentiation, preoperative CEA levels, and subgroups classified by type of resection). Results are presented as hazard ratios (HRs) and 95% confidence intervals (CIs).

Figure 1 shows the characteristics of the study cohort. Between December 1999 and December 2013, a total of 954 patients with stage IV CRC were referred to the National Cancer Center. [15] Of these, patients with histologic diagnoses other than adenocarcinoma (e.g., neuroendocrine tumor) (n = 5) and eighty-eight patients aged ≥75 years were also excluded. Ultimately, 861 patients met the aforementioned inclusion criteria, of whom 638 had been referred to the surgery division and 223 to the gastrointestinal medical oncology division.

Sixty six (8%) were adolescent and young adult patients and 795 (92%) were adult patients. Table 1 summarizes the characteristics of adolescent and young adult and adult patients. Treatment year, gender, and PS ratios did not differ between the two groups (p = 0.291, p = 0.094 and p = 0.841, respectively). Significant group-dependent differences were observed in symptoms (asymptomatic/symptomatic) (p = 0.026). Tumor location (p = 0.663), distribution of M subcategory (p = 0.784), tumor differentiation (p = 0.217), and preoperative CEA levels (p = 0.672) did not differ between the two groups.

Decisions about initial treatment were typically made by colorectal surgeons, medical oncologists, hepatobiliary surgeons, and radiologist (multidisciplinary team), who took into account disease severity and patient condition including comorbidities. Of all stage IV patients, 223 (26%) underwent resection of both the primary tumor and tumors at metastatic sites as curative resection, including 13 adolescent and young adult patients (20% of all adolescent and young adult patients) and 210 adult patients (26% of all adult patients). Resected metastatic sites in these 223 patients included the liver (n = 148), peritoneum (n = 27), lung (n = 16), and other or several sites (n = 32). Of the remaining 638 stage IV patients with unresectable tumors, 319 (50%) underwent palliative resection of the primary tumor, and 319 (50%) did not. Patients who underwent diverting stoma construction without resection of the primary tumor (n = 77), bypass surgery (n = 13), or probe laparotomy (n = 8) were included in the unresected group. In the palliative resection group, 41 patients were treated with only 5-fluorouracil, 143 patients with combination of cytotoxic agents (fluorouracil plus oxaliplatin and/or irinotecan), and 99 patients with at least one targeted agents (bevacizumab and/or anti- epidermal growth factor receptor antibody), and 2 patients with best supportive care (and missing data of 34 patients). In the unresected group, 46 patients were treated with only 5-fluorouracil, 152 patients with combination of cytotoxic agents, 94 patients with at least one targeted agents, and 22 patients with best supportive care (and missing data of 5 patients). The ratios of primary tumor resection also differed significantly between the two groups (p = 0.029) (Table 1). Specifically, adolescent and young adult patients tended not to undergo primary tumor resection, including palliative resection.

Median survival time was 13.6 months in adolescent and young adult patients and 22.4 months in adult patients, associated with the 3-year OS rates of 23.7% in adolescent and young adult patients and 33.3% in adult patients and the 5-year OS rates of 17.3% and 20.3%, respectively (p = 0.042) (Fig. 2).

Subgroup analysis was also examined by the following subgroups: patients who underwent curative resection (n = 223), patients who underwent primary tumor resection without metastasectomy (n = 319), and patients who did not undergo resection (n = 319). Figure 3 shows survival curves of young and adult patients with stage IV CRC in three subgroups. In patients who did not undergo resection (n = 349), median survival time was 10.7 months in young patients and 13.8 months in adult patients, associated with the 3-year OS rates of 3.7% in young patients and 10.5% in adult patients and the 5-year OS rates of 0% and 3.5%, respectively (p = 0.033). In contrast, prognoses of young patients and of adult patients did not differ in curative resection patients (p = 0.836) and in palliative resection patients (p = 0.983).

In univariate analysis, age was associated with prognosis (p = 0.042), and treatment year (p < 0.0001), PS (p < 0.0001), symptoms (p = 0.039), tumor location (p < 0.0001), M category (p < 0.0001), grade of liver metastasis (p < 0.0001), tumor differentiation (p < 0.0001), preoperative CEA levels (p < 0.0001), and primary tumor resection (p < 0.0001) were also significantly associated with OS (Table 2). Multivariate Cox proportional hazards regression models revealed that treatment year (1999–2004), tumor location (proximal colon), and widespread liver metastasis were independent factors associated with a worse prognosis (Table 2), whereas PS (PS 0–1), M1a subcategory, differentiated adenocarcinoma, preoperative CEA level ≤ 30 ng/ml, and curative resection were independent factors associated with better prognosis (Table 2). On the other hand, age (young versus adult) was not a significant prognostic factor.

Taken these results together, prognoses were worse in adolescent and young adult patients with stage IV colorectal cancer compared to adult patients, due to a higher proportion of patients who did not undergo resection with more advanced and severe disease, but not due to age itself.