Risk of second primary malignancies after adjuvant chemotherapy for colon cancer.

INTRODUCTION
Colorectal cancer (CRC) is among the most prevalent malignancies worldwide. Advances in adjuvant treatment have significantly improved survival rates, leading to increasing numbers of long‐term survivors.
The evolution of adjuvant chemotherapy for resected, stage II–III CRC spans 4 decades. Proof of principle arrived when the National Surgical Adjuvant Breast and Bowel C‐01 (NSABP C‐01) randomized study demonstrated that postoperative 5‐fluorouracil (5‐FU)–based chemotherapy improved both disease‐free and overall survival versus surgery alone.
1
This was soon confirmed by a North Central Cancer Treatment Group trial in which 1 year of 5‐FU plus levamisole cut the recurrence rate by 41% and cut mortality by 33% in patients with stage III CRC, establishing systemic therapy as standard for lymph node‐positive patients.
2
In further development, the NSABP C‐03 trial demonstrated that leucovorin (LV)‐modulated 5‐FU was superior to the older lomustine (MeCCNU), vincristine, and 5‐FU (MOF) regimen; whereas INT‐0089 confirmed that adding LV and shortening treatment to months preserved efficacy and reduced toxicity in high‐risk stage II/III cohorts.
3
,
4

The pooled seven‐trial analysis by Gill et al. published in 2004 quantified an absolute approximately 10% overall survival gain for 5‐FU/LV in stage III colon cancer (CC) and highlighted the marginal benefit in unselected patients with stage II disease.
5
A step change occurred with the introduction of oxaliplatin: the MOSAIC trial reported a 5.3% absolute 3‐year disease‐free survival advantage for FOLFOX4 (5‐FU/LV plus oxaliplatin) over fluoropyrimidine alone, with a durable 10‐year overall survival benefit confined largely to those with stage III disease.
6
,
7
,
8
Parallel validation came from NSABP C‐07,
9
,
10
in which bolus FLOX (5‐FU, folinic acid, oxaliplatin) achieved a 20% reduction in disease‐free survival events, and international confirmation came from the XELOXA/NO16968 trial,
11
,
12
which demonstrated that oral capecitabine plus oxaliplatin (CAPOX) matched or exceeded bolus 5‐FU/LV through 7 years of follow‐up.
However, long‐term survivors remain at risk for second and subsequent primary malignancies (SPMs), especially second colorectal tumors. A recent meta‐analysis by Du and colleagues (2022) of data from over 1.5 million CRC survivors across 42 studies reported a moderately increased risk, with a standardized incidence ratio (SIR) of 1.15 for SPMs in CRC survivors. The SIR for second CRCs was consistently increased, even when accounting for various lag periods.
13

The objectives of the current study were to assess the incidence of SPMs in CC survivors, focusing primarily on the risk of second cancers associated with different types of adjuvant chemotherapy for stage II and III CC.

MATERIALS AND METHODS

Study design
The study included all consecutive patients who had a first, primary, stage II–III CC (International Classification of Diseases codes C18 and C19) who underwent surgical treatment in the Czech Republic from 2010 to 2022. Patients who had a prior or synchronous malignancy (codes C00–C97, excluding nonmelanoma skin cancer) diagnosed within 6 months before or after the surgery were excluded. Patients with rectal cancer were not included in the current study because rectal cancer therapy often also includes radiotherapy, possibly confounding the comparison between the chemotherapy options.
Adjuvant chemotherapy was defined as treatment initiated within 3 months of the index colorectal surgery and was categorized into three groups: (1) 5‐FU or capecitabine without oxaliplatin, (2) 5‐FU or capecitabine with oxaliplatin, and (3) no chemotherapy within 3 months after surgery.

Data sources
The Czech National Cancer Registry (CNCR) was established in 1977 and is a nationwide database that records all cancer diagnoses across the entire Czech population. We used data by the year 2022. The registration of malignant neoplasms is mandated by legislation and is compulsory. The CNCR serves as a key component of the Czech National Cancer Control Program, which is tasked with providing regular and timely estimates of cancer burden in the Czech population.
The National Registry of Reimbursed Health Services has been operational since 2010 and collects comprehensive data nationwide on reimbursed services from all health insurance companies and their contracted health care providers. It includes individual payment records, data on health care service providers, personnel, technical and material resources of health care facilities, and associated metadata. The registry captures all examinations, procedures, medications, and materials that are reported and approved for reimbursement.
Both registries are components of the National Health Information System, and their data sets are fully integrated.

Statistical analysis
SIRs were calculated by dividing the observed number of SPMs and the expected number of cases. The expected numbers were derived by stratifying person‐years by sex, age (in 5‐year intervals), and calendar period (2011–2014, 2015–2018, and 2019–2022), and multiplying by the corresponding sex‐specific, age‐specific, and period‐specific cancer incidence rates from the CNCR. Exact 95% confidence intervals (CIs) for SIRs were computed under the assumption of a Poisson distribution of observed events.
Follow‐up time was defined from the date of primary surgery of the initial CRC diagnosis to the earliest of the either death or the end of follow‐up in the CNCR (December 31, 2022). Cancer cases diagnosed within 6 months before or after the surgery of the initial CC diagnosis were excluded.
The cumulative incidence function was used to estimate the cumulative probability of developing the second primary malignancy over time, accounting for competing risks. In this analysis, the event of interest was the diagnosis of an SPM after CRC, with death considered a competing risk that could preclude the occurrence of a secondary malignancy.
Statistical analyses were performed using the Stata/IC 15.1 software program (Stata Corporation).

RESULTS

Overall SPM risk
In total, 18,383 patients who had stage II and III CC were analyzed and had a follow‐up of up to 11.5 years (Table 1). Patients who received treated with fluorouracil or capecitabine plus oxaliplatin had a significantly higher overall risk of developing SPMs, with an SIR of 1.5 (95% CI, 1.4–1.6; Table 2 and Figure 1). This contrasted with those who were treated without chemotherapy or with chemotherapy excluding oxaliplatin, and both groups had an SIR of 1.1 (95% CI, 1.0–1.2). The elevated risk associated with oxaliplatin persisted when stratified by stage: in patients with stage II disease, the SIR reached 1.7 (95% CI, 1.3–2.1); and, in those with stage III disease, the SIR was 1.5 (95% CI, 1.3–1.6; see Tables S1 and S2).
The risk of developing an SPM after treatment of a first, primary CC was assessed over time using the cumulative incidence function. At 5 years, patients who received with fluorouracil or capecitabine combined with oxaliplatin had an 8.4% risk (95% CI, 7.5%–9.3% risk) of developing an SPM compared with a 7.9% risk (95% CI, 7.0%–8.7% risk) in those treated without oxaliplatin and a 7.8% risk (95% CI, 7.2%–8.4% risk) in patients who did not receive chemotherapy (Table 2). This difference became more pronounced at 10 years, with cumulative incidence rates of 14.6% (95% CI, 13.1%–16.1%), 12.5% (95% CI, 11.3%–13.7%), and 11.0% (95% CI, 10.1%–11.8%) for the fluorouracil plus oxaliplatin, capecitabine plus oxaliplatin, and no‐chemotherapy groups, respectively (Figure 1).

Risk of subsequent colorectal cancer
A detailed, site‐specific analysis revealed particularly significant increases in the SIR for second and subsequent SPMs. Patients receiving oxaliplatin‐based chemotherapy had a significantly higher second CRC risk (SIR, 2.2; 95% CI, 1.8–2.7) compared with those who received fluorouracil/capecitabine alone (SIR, 1.3; 95% CI, 1.0–1.6) or no chemotherapy (SIR, 1.5; 95% CI, 1.3–1.8; Table 2 and Figure 2). This was consistent when analyzing patients with stage II or stage III disease separately. Patients who received oxaliplatin had significantly elevated SIRs (stage II: SIR, 1.8; 95% CI, 0.9–3.2; stage III: SIR, 2.3; 95% CI, 1.8–2.9; see Tables S1 and S2). These values exceeded those observed in patients who did not receive chemotherapy (stage II: SIR, 1.4; 95% CI, 1.1–1.7; stage III: SIR, 2.0; 95% CI, 1.5–2.8). The SIR for second CRC in patients who received fluorouracil or capecitabine without oxaliplatin was only moderately elevated. For patients with stage II disease, the SIR was 1.4 (95% CI, 1.1–1.9); and, for those with stage III disease, the SIR was 1.1 (95% CI, 0.8–1.6).
At 5 years, the risk ranged from 1.4% (95% CI, 1.1%–1.8%) in patients treated who received fluorouracil or capecitabine without oxaliplatin to 1.7% both in patients who received oxaliplatin (95% CI, 1.3%–2.1%) and in those who did not receive chemotherapy (95% CI, 1.4%–2.0%). By 10 years, the risk of a second colorectal cancer rose to 2.8% (95% CI, 2.1%–3.5%) in the oxaliplatin group versus 2.0% in those treated without oxaliplatin (95% CI, 1.5%–2.5%) and in those treated without chemotherapy (95% CI, 1.6%–2.3%; Figure 2).

Risk of subsequent noncolorectal cancers
For lung cancer, the SIR rose to 1.6 (95% CI, 1.2–2.0) among patients with stage III disease who received oxaliplatin. In comparison, the SIRs remained below 1.0 in the other treatment groups. For bladder cancer, a striking increase was identified in patients who had stage II disease treated with oxaliplatin, with an SIR of 3.3 (95% CI, 1.3–6.8), in contrast to SIRs between 0.9 (95% CI, 0.6–1.3) in the group who received no chemotherapy and 1.1 (95% CI, 0.6–1.9) in the group that received fluorouracil or capecitabine without oxaliplatin (Table 2; see Tables S1 and S2).
Regarding prostate cancer, patients who received fluorouracil or capecitabine with oxaliplatin demonstrated a moderately elevated SIR, particularly in stage II CC survivors. Conversely, patients who were treated without chemotherapy exhibited lower or average SIRs for prostate cancer (stage II; SIR, 0.9; 95% CI: 0.7–1.1; stage III: SIR, 0.8; 95% CI, 0.5–1.3). For those who received fluorouracil or capecitabine without oxaliplatin, the SIRs were 1.1 (95% CI, 0.8–1.5) in patients with stage II prostate cancer and 0.9 (95% CI, 0.6–1.3) in those with stage III prostate cancer (Table 2; see Tables S1 and S2).
Among patients who had thyroid cancer, the SIR was 3.0 (95% CI, 1.7–5.1) for those who did not receive chemotherapy, whereas the risk was not increased compared with the general population for the chemotherapy groups, although the number of patients was relatively low (Table 2; see Tables S1 and S2).
Moreover, the category of other cancers demonstrated consistently increased SIRs across all groups but the SIR was particularly elevated in patients with oxaliplatin exposure. In patients with other stage III cancers, the SIR reached 3.5 (95% CI, 2.6–4.6), whereas those with other stage II cancers had SIRs of 3.0 (95% CI, 1.3–5.9).
Conversely, some cancer types exhibited no meaningful elevation in the SIR across treatment groups. These included noncolorectal gastrointestinal cancers, such as esophageal, gastric, pancreatic, biliary, and liver cancers.
Patients who had nonmelanoma skin cancer demonstrated a reduced risk in the fluoropyrimidine‐only group and in the group that not receive chemotherapy, whereas the risk for the group that received oxaliplatin corresponded to the risk in the general population.

DISCUSSION
Standard postoperative management of CC includes fluoropyrimidine‐based chemotherapy with or without oxaliplatin or observation. This variability of treatment options provides an opportunity to study the effect of various chemotherapy drugs in long‐term survivors, for whom the risk of second cancers is emerging as a mayor health concern.
Based on the population‐level data provided, the risk of developing SPMs varied significantly, depending on the chemotherapy regimen used after treatment of the initial CC. Patients who received fluorouracil or capecitabine combined with oxaliplatin demonstrated the highest overall risk of developing SPMs compared with other categories. The most pronounced elevations in risk were observed for colorectal, prostate, and lung cancers.
The risk of developing second primary CRCs varied significantly, depending on the chemotherapy regimen used after treatment of the initial CC. Among patients with stage II and III CC who received fluorouracil or capecitabine with oxaliplatin, the SIR for a second CRC was elevated. In contrast, fluoropyrimidine‐only chemotherapy was not associated with an increased risk of a SPM. This trend was consistent across both stage II and II cancer, indicating that the effect of oxaliplatin may be independent of the extent of initial disease. Among the patients who did not receive chemotherapy, the SIR for second CRCs also was elevated, particularly in patients with stage III disease, suggesting that patient selection and baseline risk also may influence outcomes. However, the lower SIR observed in patients with stage II disease in the nonchemotherapy group indicates that oxaliplatin‐based chemotherapy remains a relevant variable in assessing the risk of secondary cancer. In addition, the risk of lung and prostate cancer was also elevated in patients who received oxaliplatin. These findings support the hypothesis that oxaliplatin may be a significant contributing factor to long‐term secondary colorectal carcinogenesis. In contrast, as indicated in Figure 2, the risk of second primary CRC appears lower in both chemotherapy cohorts compared with the nonchemotherapy cohort during the first 3 years, possibly indicating the eradication of some initial neoplasms by adjuvant chemotherapy.
Oxaliplatin plays a crucial role in CRC treatment, yet its long‐term effects on the tumor microenvironment and patient outcomes remain significant concerns and have been the subject of several recent studies. A pivotal study by Linares et al. in 2023 demonstrated that oxaliplatin persists in cancer‐associated fibroblasts long after treatment cessation, triggering the transforming growth factor‐beta signaling pathway. This activation leads to the secretion of protumorigenic factors, such as periostin, that enhance tumor aggressiveness and chemotherapy resistance. It is plausible that these fibroblasts may promote the growth of second CRCs.
14
Stojanovska et al. demonstrated that oxaliplatin treatment increases HMGB1 (high‐mobility group box 1) expression and induces morphologic changes in TLR4‐positive immune cells within the colonic lamina propria. Furthermore, oxaliplatin alters the gut microbiota at the genus level, significantly reducing Parabacteroides and Prevotella1 while increasing Prevotella2 and Odoribacter. The authors also reported suppressed expression of TLR7, TLR9, and the major histocompatibility class I gene H2‐D1 as well as reductions in macrophage and dendritic cell populations in mesenteric lymph nodes. These findings suggest that oxaliplatin induces immunosuppressive and microbiota‐altering effects in the colon, mediated by oxidized damage‐associated molecular patterns and altered antigen‐presentation mechanisms.
15
Furthermore, there is evidence that chronic exposure of CRC cell lines to oxaliplatin induces epithelial‐to‐mesenchymal transition, a process associated with increased cellular motility and invasiveness.
16

Several studies have explored the association between chemotherapy and SPM risk in CRC in addition to the proven risk factors for SPMs that include older age, male sex, localized or regional stage at diagnosis, treatment with surgery, and grade 1 or 2.
17
Except for the first two, these factors reflect the probability of long‐term survival after the first CRC, allowing sufficient time for an SPM to develop. In a large Surveillance, Epidemiology, and End Results (SEER)‐based study conducted by Yang et al. in 2023, patients who received chemotherapy had a moderately increased risk of SPMs compared with patients who did not receive chemotherapy (subdistribution hazard ratio, 1.10; 95% CI, 1.06–1.14; p < .01).
18
In another SEER‐based study, Gao et al. (2024) explored the association between chemotherapy for rectal cancer and the risk of developing second primary endometrial cancer. Chemotherapy and radiotherapy recipients had a significantly higher risk, with a hazard ratio for second primary endometrial cancer in patients who had chemotherapy‐treated RC of 2.26, suggesting chemotherapy as an independent risk factor.
19

However, the SEER registry data do not enable analyses of various chemotherapy regimens, and there are few studies specifically addressing the risk associated with different cytostatic drugs. Green and collaborators, analyzing 3278 patients enrolled in the Intergroup 0089 trial of four different regimens of adjuvant 5‐FU–based chemotherapy, observed a 5‐year cumulative incidence of 1.5% for second colon cancers, with a corresponding SIR of 1.6 compared with the general population.
20
Teufel et al. also focused on the risk of secondary cancers in patients who received adjuvant chemotherapy from a cohort of 2856 patients with mostly stage III CC. Patients who received FOLFOX (folinic acid [LV], 5‐FU, and oxaliplatin) had a lower SPM risk than those who received FUFOL (5‐FU/LV; 6.1% vs. 13.9%, respectively), although the result lacked statistical significance in multivariable analysis. However, the inclusion of patients who had stage IV disease (10% of the total), in which the first cancer is expected to be the dominant cause of mortality, may weaken the conclusions of this analysis. That study demonstrated no significant increase in the overall risk of second cancer risk among chemotherapy‐treated patients, analyzing hazard ratios of chemotherapy‐treated versus chemotherapy‐untreated patients, in contrast to our study, in which we used the SIR (i.e., the adjusted comparison with the general population).
21

Our data indicate that, in addition to an increased risk of second CRC, oxaliplatin was also associated with an increased risk of lung and prostate cancers. There are no reports of this phenomenon in the published literature, and the association needs further study. Current research does not suggest that the receipt of oxaliplatin is associated with a risk of skin cancer, which was decreased in the our study.
A key strength of the current study lies in its large cohort size of greater than 18,000 patients and a follow‐up period up to 11.5 years, which confer strong statistical power even when analyzing the relevant disease stages separately and allow for the observation of long‐term trends. The study population included patients without systemic adjuvant treatment, providing internal control. Moreover, the study's design reflects contemporary clinical practice, with patients stratified by treatment modality—namely, surgery alone, chemotherapy with fluorouracil/capecitabine, or combination regimens that included oxaliplatin—and is the first study to our knowledge to separately analyze the types of chemotherapy. Calculation of the SIR inherently includes adjustment for age and sex, which are the only two established risk factors for SPMs in CC.
Despite its contributions, the current study is not without limitations. As a retrospective cohort analysis, it is subject to inherent risks of selection bias and residual confounding. Although adjustments were made for demographic variables and treatment periods, other relevant factors—including genetic predisposition, lifestyle behaviors, comorbidities, and socioeconomic status—were not fully accounted for because of the lack of relevant data in the registries, and this may have influenced the observed outcomes. Furthermore, the study does not report on treatment‐specific variables, such as chemotherapy dose intensity or duration, which could substantially affect the risk of secondary malignancy.
In conclusion, CC survivors, especially those treated with oxaliplatin‐based chemotherapy, are at heightened risk for SPMs. These findings emphasize the need to weigh the benefits of oxaliplatin‐based regimens against their long‐term implications. Risk stratification and individualized survivorship planning should be considered, incorporating both the initial cancer characteristics and the treatment history to optimize follow‐up and screening for metachronous tumors.

AUTHOR CONTRIBUTIONS

Tomas Buchler: Conceptualization; investigation; writing—original draft; validation; writing—review and editing; project administration; supervision. Monika Ambrozova: Conceptualization; investigation; writing—review and editing; data curation; formal analysis; methodology. Ondrej Majek: Conceptualization; investigation; writing—review and editing; formal analysis; data curation; methodology. Tereza Dianova: Conceptualization; investigation; methodology; writing—review and editing; formal analysis; data curation. Petr Klika: Conceptualization; investigation; methodology; writing—review and editing; formal analysis; data curation. Ladislav Dusek: Conceptualization; writing—review and editing; validation; project administration; supervision; resources.

CONFLICT OF INTEREST STATEMENT
Tomas Buchler reports personal/consulting and/or advisory fees or honoraria from Astellas Pharma, AstraZeneca, Bristol Myers Squibb Company, Eli Lilly and Company, Ipsen, Johnson & Johnson Health Care Systems Inc., Merck, Merck Sharp & Dohme, Novartis, Pfizer, and Roche outside the submitted work. The remaining authors disclosed no conflicts of interest.

Supporting information
Table S1

Table S2