Pre-Surgery Inflammatory and Angiogenesis Biomarkers as Predictors of 12-Month Cancer-Related Distress: Results from the ColoCare Study

Colorectal cancer is the third most diagnosed cancer in both men and women in the United States and is estimated to account for 149,500 new diagnoses and 52,580 deaths (1, 2). Although improvement of cancer screenings and treatments has resulted in increased survival rates, patients with colorectal cancer are likely to suffer from distress, including depression and anxiety, as a result of the disease process and cancer treatment (3). The National Comprehensive Cancer Network (NCCN) defined cancer-related distress as a “multi-factorial unpleasant emotional experience of psychological (cognitive, behavioral, and emotional), social, and/or spiritual nature that may interfere with the ability to cope effectively with cancer, its physical symptoms and its treatment” (4). When cancer-related distress is unrecognized or undertreated, patients may experience impaired decision making, decreased satisfaction, depression, anxiety, isolation, panic, and existential and spiritual crisis (5).

Underlying mechanisms of cancer-related distress remain unclear, but an understanding of these mechanisms is essential to develop targeted interventions to reduce distress in this vulnerable population. Systemic inflammation, a hallmark of cancer, is one hypothesized mechanism as it has been associated with psychological distress, including depression (6–8). Consistent evidence indicates that antidepressant drugs possess anti-inflammatory properties (9–12). A randomized clinical trial with n = 241 patients with major depressive disorder, concluded that lower C-reactive protein (CRP) concentration at enrollment predicted greater improvement of depressive symptoms after 12 weeks (P < 0.05), suggesting that inflammation plays a role in the progression of depression (12).

Cancer-related distress can have a complex relationship with symptoms that may influence clinical outcomes. We have previously shown that younger age, advanced cancer stage, lower income, lower social support levels, and worse functional status and low quality of life pre-surgery significantly predicted higher cancer-related distress after surgery (13). A pooling of data from 16 prospective cohort studies, including n = 163,363 patients with colorectal cancer, demonstrated that greater levels of distress led to higher risk of cancer mortality (14). Another study reported that surgery-induced stress is a powerful factor promoting malignant tumor growth involving inflammation, sympathetic nervous system activation, and increased cytokine release, including angiogenesis pathways such as VEGF, IL6, and IL8 that increased risk of cancer recurrence (15). In short, there may be important associations between biomarkers of inflammation and angiogenesis with both psychological and physical symptoms in colorectal cancer survivors.

Prior studies reported a positive association between elevated proinflammatory biomarkers and behavioral symptoms in the general population. Studies have revealed elevated levels of proinflammatory biomarkers, including IL6, TNFα, and CRP in patients with increased anxiety (16). A longitudinal study revealed that CRP is a risk marker for development of depression 10 years after follow-up in women who had no history of depression at enrollment, independent of factors, including adiposity, weight, and body mass index (BMI; ref. 6). Another large-scale prospective study recruited n = 10,357 participants with no history of depression and revealed that inflammation was predictive of depression 5 years later (7). Several other longitudinal studies have revealed this directionality in the general population, which may also be present in patients with colorectal cancer (8, 17).

Although studies have investigated the link between inflammation and distress in cancer populations, research is limited for colorectal cancer (18–20). A cross-sectional study correlated stress with higher levels of TNFα in patients with chronic lymphocytic leukemia (19). Another cross-sectional study showed that different stages of antitumor therapy revealed positive correlations between anxiety and depression with IL6, IL8, and TNFα (18). Research suggests that inflammation, common before and during treatment for patients with cancer, interacts with distress; however, prospective research is still lacking.

We have previously shown, that in addition to inflammation, elevated angiogenesis biomarkers are associated with poorer clinical outcomes in patients with colorectal cancer; however, limited research has linked angiogenesis biomarkers to cancer-related distress (21, 22). In 2006, an in vivo study in mice reported significantly increased tumor vascularization and enhanced expression of VEGF in experimentally distressed animals using a physical restraint system designed to induce nervous system activity characteristic of chronic stress (23). In 2002, a cross-sectional study revealed that women with ovarian carcinoma who reported higher levels of social support had lower concentrations of VEGF (24). There is evidence that angiogenesis is associated with distress; however, there is a lack in prospective research in patients with colorectal cancer (22–24).

On the basis of prior findings of relationships between inflammatory and angiogenesis processes and stress, this study investigates associations between serum concentrations of pre-surgery and 12 months post-surgery inflammatory and angiogenesis-related biomarkers [CRP, serum-amyloid A (SAA), IL6, IL8, VEGF-A, VEGF-D, TNFα, soluble intercellular adhesion molecule-1 (sICAM-1) and soluble vascular cell adhesion molecule-1 (sVCAM-1)] with cancer-related distress 12 months after surgery. We hypothesized that higher pre-surgery inflammation and angiogenesis would be associated with elevated distress at 12-months. We further expected that higher biomarker concentration for inflammation and angiogenesis at the 12-month time point would be associated with elevated cancer-related distress at the 12-month time point.

This study includes patients from two study centers enrolling in the international prospective ColoCare Study (ClinicalTrials.gov Identifier: NCT02328677; ref. 25). The ColoCare Study includes men and women ages 18–89 years who were newly diagnosed with stage I–IV colon, rectum, or rectosigmoid cancer before surgery and were mentally and physically able to consent and participate in the study (13, 25–27). For the present study, patients who had previously been recruited were newly selected given the eligibility criteria. These patients were selected at the National Center for Tumor Diseases (Heidelberg, Germany), and the Huntsman Cancer Institute (HCI; Salt Lake City, UT).

For the present study, patients from HCI and HD with measured inflammatory and angiogenesis biomarkers at pre-surgery and at 12-months follow-up and who were administered the CTXD score at 12-months follow-up were included in the analyses. All analyses in this article are based on data and biospecimens collected between 2010 and 2019. This study was approved by the ethics committees of the Medical Faculty at the University of Heidelberg and the University of Utah and was conducted in accordance with the 1964 Helsinki Declaration. All subjects provided written informed consent.

Pre-surgery blood samples were collected from non-fasting patients before surgery at HCI and HD. Serum from patients at HCI was extracted from whole blood and stored at −80°C in 110 μL aliquots until analysis. Serum samples from patients in Heidelberg were extracted within 4 hours of blood draw, stored in 500 μL aliquots at −80°C and shipped on dry ice to HCI for analysis.

Serum-based assays for multiplexed IL6, IL8, TNFα, CRP, SAA, sICAM-1, sVCAM-1, VEGF-D, and VEGF-A have previously been established on the Mesoscale Discovery Platform (MSD) in the Ulrich laboratory at HCI (21, 27). The biomarker panel was selected on the basis of previous studies that investigated central biomarkers for inflammation and angiogenesis and behavioral outcomes in both the general and cancer population (12, 18, 22, 26–28).

All samples were blinded, run in duplicate, and read on an MSD Sector Imager 2400A. For both sites, VEGF-A and VEGF-D were run on a V-Plex Angiogenesis Panel Human Kit at a dilution of 1:8. For the HCI samples, IL6, IL8, and TNFα were run on a V-Plex Proinflammatory Panel 1 Human Kit at a dilution of 1:2 and CRP, SAA, sICAM-1 and sVCAM-1 were run on a V-Plex Vascular Injury Panel 2 Human Kit at a dilution of 1:2,500. For HD, CRP, SAA, sICAM-1 and sVCAM-1 were run on a V-Plex Vascular Injury Panel 2 Human Kit at a dilution of 1:1,000. IL6, IL8, and TNFα were run on a custom U-PLEX Proinflammatory Panel at 1:2 and previously reported with lower levels (26, 27). The data were subsequently bridged to V-PLEX Proinflammatory Panel 1 levels. Three quality controls were run on each plate and used to calculate the inter- and intra-plate coefficient of variabilities (CV). The overall inter-plate CV for HD and HCI samples was 9.9% and 9.6% and the intra-plate CV was 4.6% and 3.9%, respectively.

Cancer-related distress was evaluated by the Cancer and Treatment Distress (CTXD) score designed to measure worry or distress related to cancer and its treatment (28). The CTXD score has demonstrated sensitivity and specificity to cancer distress, and elevated scores capture nearly all patients with clinical depression and anxiety symptoms in patients with cancer (13). The original measure includes 22 items and 5 subscales: Uncertainty, family strain, health burden, finances, and medical demands (13, 29, 30). For this study, to reduce responder burden, we used a brief 13-item version that correlates highly with the 22-item version (r = 0.99). The 13-item version included 4 subscales: Uncertainty, family strain, health burden, and medical and financial demands, each of which correlates with the longer subscales (r = 0.88–0.97). Internal consistency reliability was strong for the overall scale (α = 0.92) as well as for subscales (α = 0.88 to α = 0.78). The CTXD score has demonstrated sensitivity and specificity to cancer distress and is strongly associated with depression, anxiety, and post-traumatic stress symptoms in patients with cancer (31–33). Each item (e.g., including the overall scale) was scored on a scale ranging from 0 = “none” to 3 = “severe” to determine distress or worry in the past week with overall and subscales calculated as mean scores. The overall score and all subscores were dichotomized into high (≥0.90) and low distress (<0.90) based on evidence that this score captures 90% of those with symptoms of clinical depression or anxiety (29, 33, 34).

Pre-surgery and 12-month biomarker levels, as well as 12-month CTXD scores were included in the primary analysis. Patients who had at least one biomarker measurement and who have completed the CTXD questionnaire were included in the analysis. Biomarker levels were log₂-transformed to prevent heteroscedasticity. Potential confounding was investigated for age (years), sex (male, female), BMI (kg/m²), tumor stage (I–IV), tumor site (colon, rectal), adjuvant treatment (yes, no), regular NSAID/aspirin drugs [NSAIDs use at least 1 time per week during the past month (yes, no)], and study center (HCI, HD). The final model was adjusted for covariates of age, sex, BMI, tumor stage, and study center. We further adjusted analysis at the 12-month time point for baseline biomarker levels.

Mean and standard deviation (SD) were calculated for age, BMI, biomarker concentrations and CTXD scores. Frequency and percentages were calculated for sex, tumor stage, neoadjuvant and adjuvant treatment, surgery (yes, no), tumor site (colon, rectal), and study center.

To investigate associations between pro-inflammatory and angiogenic biomarkers and CTXD scores, Pearson's partial correlations were conducted. Correlation coefficients were calculated between pre-surgery biomarker concentrations with CTXD scores, and 12 months post-surgery biomarker concentration with CTXD scores. The 12-months analyses were additionally adjusted for baseline biomarker concentrations. To further investigate the effect of the biomarkers, we conducted logistic regressions to evaluate pre-surgery biomarkers as predictors of elevated cancer-related distress 12 months after surgery. Odds ratios (OR) and 95% confidence intervals (CI) were calculated to determine the associations of doubling of biomarker concentration with risk of high distress. Additional stratified analyses were conducted to determine effect modification by tumor stage (metastatic, non-metastatic) and by study center (HCI, HD). We tested for heterogeneity between study centers to ensure that combined analyses, including patients from both sites, were valid. We performed tertile analyses to confirm whether biomarkers had a linear relationship with the CTXD scores.

Statistical analyses were performed using SAS Studio OnDemand for Academics (SAS Institute Inc.), for Google Chrome 27+, version 3.8. All tests were two-sided and considered statistically significant when P < 0.05.

The ColoCare Study data are available from [email protected] on reasonable request and as described previously on the ColoCare website (https://uofuhealth.utah.edu/huntsman/labs/colocare-consortium/). Our data sharing procedures have been updated and are available online (https://uofuhealth.utah.edu/huntsman/labs/colocare-consortium/data-sharing/new-projects.php). For any additional questions please contact the ColoCare Study Administrator Team ([email protected]).

Table 1 summarizes the patient characteristics at baseline (pre-surgery). The mean age of patients was 62 years and mean BMI was 28 kg/m². Patients were predominantly male (63%). Patient characteristics stratified by study center are presented in Supplementary Table S1. The study centers had similar rates of adjuvant treatment, comparing HCI and HD (44% vs. 45%) and neoadjuvant treatment comparing HCI and HD (27% vs. 33%). Patients in HD had lower BMI than patients from HCI (27 kg/m² vs. 30 kg/m²; P < 0.01) and were older than patients from HCI (62 vs. 61; P = 0.37).

The geometric means of pro-inflammatory and angiogenic biomarkers are presented in Table 2 and separated by study center in Supplementary Table S2. Pre-surgery, significant differences between centers were observed for CRP, sICAM-1, IL6, and IL8 (P < 0.01, P = 0.02, P = 0.02, P < 0.01, respectively). 12 months after surgery, significant differences between centers were observed for CRP, SAA, and sVCAM-1 (P < 0.01, P < 0.01, P < 0.01, respectively). [Note, IL6, IL8, and TNFα were not measured in HCI samples at 12-months].

Overall mean CTXD scores are shown in Table 3 and scores stratified by study center are shown in Supplementary Table S3. Among the 4 subscales, the mean distress for uncertainty was the highest at both study centers. The lowest distress for HCI was family strain, and the lowest distress for HD was managing medical and financial demands. For each subscore and the overall score, patients in the US reported greater distress than in Germany. Significant differences in CTXD-defined distress between sites were observed for the overall score, uncertainty, health burden, and managing medical and financial demands (P < 0.01, P = 0.02, P = 0.02, P < 0.01, respectively).

Pearson partial correlations are presented in Figs. 1 and 2. A positive correlation coefficient (r) indicates that increased biomarker concentrations are correlated with distress and an inverse correlation coefficient indicates that increased biomarker concentrations are correlated with decreased distress. Any correlation with a P value of <0.05 was considered statistically significant. We performed Bonferroni adjustment for multiple testing for each of the investigated subscales. Logistic regressions adjusted for age, sex, BMI, tumor stage, and study center are presented in Table 4. Additional analyses stratified by study center are presented in Supplementary Figs. S1–S2 and Supplementary Table S4.

Pre-surgery IL6, VEGF-A, and VEGF-D levels were correlated with an increase in overall CTXD score (r = 0.14; P = 0.03, r = 0.14; P = 0.03, r = 0.16; P = 0.02, respectively).

Twelve months post-surgery, these correlations were no longer significant; however, IL6, and VEGF-D correlations were in a similar direction (r = 0.10 P = 0.46, r = 0.04; P = 0.45, respectively).

Doubling of pre-surgery IL6 concentration predicted a 20% increased risk of having high distress (OR, 1.20; 95% CI, 1.02–1.41; P = 0.03) 12 months after surgery.

Uncertainty was the highest observed subscore of the CTXD score across all patients and was endorsed by half of the survivors (52% with scores >0.90). No biomarkers were statistically significantly correlated with uncertainty at either time point. Pre-surgery IL6 was marginally significantly associated with a 16% increased risk of high uncertainty (OR, 1.16; 95% CI, 0.98–1.38; P = 0.08).

Pre-surgery VEGF-A and VEGF-D concentrations were correlated with increased family strain (r = 0.15; P = 0.02, r = 0.20; P < 0.01, respectively). Twelve months after surgery, IL6 was correlated with family strain (r = 0.25; P = 0.01). The associations of VEGF-A and VEGF-D with family strain were significant in the logistic regression model. Doubling of pre-surgery VEGF-A concentration predicted a 21% increased risk of high family strain (OR, 1.21; 95% CI, 1.01–1.44; P = 0.04). Doubling of pre-surgery VEGF-D predicted a 45% increased risk of high family strain (OR, 1.45; 95% CI, 1.06–1.99; P = 0.02).

Pre-surgery IL6, TNFα, and VEGF-A concentrations were correlated with increased health burden (r = 0.17; P = 0.01, r = 0.14; P = 0.04, r = 0.13; P = 0.04, respectively). No biomarkers measured 12 months post-surgery were associated with health burden. Pre-surgery IL6 also predicted a 26% increased risk in health burden (OR, 1.26; 95% CI, 1.06–1.49; P = 0.01).

Pre-surgery VEGF-A and VEGF-D were correlated with increased medical and financial demands (r = 0.13; P = 0.04, r = 0.17; P = 0.01, respectively). However, 12 months after surgery, the biomarkers were no longer associated.

Pre-surgery VEGF-D predicted a 34% increased risk of medical and financial demands (OR, 1.34; 95% CI, 1.01–1.74; P = 0.03).

A test of heterogeneity in the associations showed comparable results across study centers, except for the association of pre-surgery VEGF-A with the overall score and the health burden score (P_{heterogeneity} = 0.02, P_{heterogeneity} = 0.03). Notably, VEGF-A was not significantly associated with the overall score or health burden in any of our statistical analyses. We did not observe heterogeneity in the associations of any biomarker with CTXD score in logistic regression models as shown in Supplementary Table S5 (P_{heterogeneity} > 0.05). We additionally performed analyses stratified by tumor stage, neoadjuvant treatment, adjuvant treatment, and NSAID use. None of these variables were included in the final model. We further tested for linearity in the associations and investigated the P for trend in tertile analyses. Pre-surgery biomarkers had a dose–response relationship with distress outcomes, as shown by the tertile analyses in Supplementary Table S6.

In this international, prospective study, the angiogenesis biomarkers VEGF-A and VEGF-D, were associated with distress in patients with colorectal cancer. VEGF-A and VEGF-D were correlated with increased overall distress, family strain, health burden (VEGF-D only) and medical and financial demands at the 12-month time point. Results from the logistic regression analyses showed that pre-surgery VEGF-A and VEGF-D concentrations predicted high family strain, and high medical and financial demands (VEGF-D only) 12 months after surgery. IL6 was concurrently associated with overall cancer-related distress and subscales at pre-surgery and 12 months. These findings provide first evidence that angiogenesis and inflammation are associated with cancer-related distress 12 months after surgery. Although the prediction is directional, the causality may be reverse. However, the association between pre-surgery biomarker levels and distress continues to demonstrate an impact at 12 months. On the basis of sensitivity tests, we conclude that the measured biomarkers were associated with distress outcomes independent of disease factors, including tumor stage, tumor treatment, and study center. Angiogenesis may particularly be associated with socioeconomic and stress disparities such as financial demands and family strain (35, 36).

The associations between angiogenesis and cancer-related distress are consistent with a cross-sectional study that revealed women with ovarian cancer who reported high levels of social support had lower levels of pre-surgery VEGF (24). Our results showed that pre-surgery inflammation and angiogenesis biomarkers were associated with distress 12 months post-surgery. Longitudinal data of n = 10,357 cancer-free adult participants with no evidence of depression at enrollment, revealed that increased CRP and white blood cell count (indicators of inflammation) levels at enrollment, predicted new cases of depression 5 years later (7). Another study in the general population found that total white blood cell count is linked to an increase in depressive symptoms without evidence of a bi-directional relationship (17). In a systemic review on the effects of inflammation on depression, it was shown that if the peripheral immune system continues unabated, such as during systemic infections or cancer, the immune signaling can lead to exacerbation of symptoms and to the development of depression (8). The authors conclude that inflammation is an important biological event that might increase the risk of major depressive episodes, much like the more traditional psychosocial factors. Although the above-mentioned studies investigate inflammation and distress in the general population, the directionality is consistent with our results in patients with colorectal cancer.

Pre-surgery CRP, an acute inflammation biomarker that has been associated with distress, was not associated with distress 12 months after surgery in our study (6, 7, 16). Acute inflammation, such as increased CRP levels, plays a more immediate role in the acute setting and does not predict distress one year after inflammation was measured.

Interestingly, increased concentrations of pre-surgery angiogenesis biomarkers are correlated with family strain, whereas 12 months post-surgery concentrations of other acute inflammation biomarkers are correlated with family strain. These results suggest that angiogenesis biomarkers are correlated with a longer-term or delayed distress response.

Cancer-related distress has well known impacts on quality of life in patients and long-term survivors. In response to this established fact, the NCCN Distress Management and the NCCN Survivorship Guidelines each provide detailed recommendations for the assessment and treatment of cancer related distress, depression, anxiety, and post-traumatic stress symptoms (37). Similarly, the American Cancer Society guidelines specific to colorectal cancer survivors recommend at least annual surveillance and treatment of elevated distress (38). The National Cancer Institute provides guidance for health care professionals in screening and treating distress (39). Our results confirm that distress not only interacts with other quality of life outcomes, but also with biologic factors.

We have previously reported that pre-surgery angiogenesis may also affect long-term survival (21). Pre-surgery VEGF-D has been associated with worse overall survival in patients with colon cancer (21), and a 3-fold increased risk of death in patients with rectal cancer (21, 40). The present study demonstrates that pre-surgery angiogenesis has an impact on cancer-related distress. In ongoing research, we are examining whether postoperative distress may interact with long-term survival outcomes.

Our findings showed that patients had increased distress at our US-based site (HCI) compared with our Germany-based site (HD), specifically in the subscale's uncertainty, health burden, and medical and financial demands. CTXD score averages, correlation coefficients, and ORs were generally stronger at HCI. The differences between study centers may be due to differences in health care systems between the two countries. Patients in the US are responsible for co-pays and high out-of-pocket expenses, which are nearly non-existent for patients in Germany, who are covered by universal health care (40). One study using the CTXD score to determine risk factors for cancer-related distress in colorectal cancer survivors revealed that stable economic status pre-surgery is an important factor in preventing cancer-related distress over time (13). These differences may give rise to broad-based stress as seen with the higher levels of uncertainty, health burden, and medical and financial demands. The results from the heterogeneity test and the adjustment by study center in our analyses assures that the findings are applicable to patients across the study centers.

This study has several strengths and limitations. This study included a prospective longitudinal design, using a validated measure, CTXD score, to assess the association between inflammation and angiogenesis biomarkers with cancer-related distress. Because distress was not measured at baseline using the CTXD score, we cannot rule out ongoing associations of inflammation and distress throughout the course of treatment versus a direct effect of inflammation on distress. We used the electronic data warehouse at HCI to identify patients with a history of depression before their cancer diagnosis. We adjusted the main model for baseline depression; however, no changes were observed in the presented ORs in comparison with the ORs before and after adjustment for depression. Unfortunately, these data are not available for the German cohort. Although our results show the prediction is directional, from pre-surgery to 12 months, the association may be bidirectional. The critical point here is that biobehavioral processes do interact and may be amenable to biobehavioral interventions. The study design limits the risk of reverse causality and allows us to draw direct conclusions. This study used validated subscales, adding specificity to the assessment. We did not assess pre-surgery cancer-related distress and could not control for pre-surgery distress in our analyses. IL6, IL8, and TNFα data were only available pre-surgery for HCI patients, lowering the statistical power for these biomarkers. The magnitude of each correlation coefficient is relatively weak and may be due to small sample size. We do not have pathologic or molecular information of inflammatory indicators in primary tumors. A recent Nature publication highlights the interaction between systemic inflammation markers and primary tumor inflammation status. Zhao and colleagues (41) report that acute, primary tumor inflammation acts as a defense mechanism against infection or injury that can regulate antitumor immune responses. However, if the acute inflammatory reaction remains unresolved, it can transform into chronic inflammation generating an immunosuppressive tumor-promoting environment. Our findings contribute evidence that this process also may interact with stress response factors reflected in perceived distress.

This is the first study to investigate a large comprehensive panel of inflammation and angiogenesis-related biomarkers in associations with cancer-related distress 12-months after surgery. In the present study, inflammation and angiogenesis biomarkers measured at 12 months are both increased by distress and increase with distress. The underlying biology for these predicted effects remains to be illuminated through additional research. Importantly, this is the first study to show that pre-surgery angiogenesis biomarkers are correlated with and may predict cancer-related distress 12 months after surgery in patients with colorectal cancer, though the underlying mechanism is unknown. Previous studies have revealed distress as a predictor of inflammation, whereas the present study shows that inflammation and angiogenesis can predict cancer-treatment–related distress 12 months after surgery.

In conclusion, pre-surgery inflammation and angiogenesis-related biomarkers are associated with cancer-related distress 12 months after surgery. On the basis of our findings, interventions focused on lowering inflammation and angiogenesis, such as diet, exercise, and anti-inflammatory medication, may improve cancer-related distress and quality of life in patients with colorectal cancer.

B. Gigic reports grants from NCI and BMBF during the conduct of the study. C.A. Warby reports grants from NIH and grants, non-financial support, and other support from Huntsman Cancer Foundation during the conduct of the study; as well as grants from NIH, grants and other support from Huntsman cancer foundation, non-financial support from Huntsman Cancer Institute, and grants from University of Utah III outside the submitted work. E.M. Siegel reports grants from NCI during the conduct of the study. C.M. Ulrich reports grants from NIH and Huntsman Cancer Foundation during the conduct of the study. No disclosures were reported by the other authors.

C.L. Lindley: Data curation, formal analysis, writing–original draft, writing–review and editing. B. Gigic: Resources, data curation, formal analysis, investigation, writing–original draft, writing–review and editing. A.R. Peoples: Resources, data curation, investigation, writing–review and editing. C.J. Han: Methodology, writing–review and editing. T. Lin: Formal analysis, visualization, methodology, writing–review and editing. C. Himbert: Data curation, formal analysis, visualization, writing–review and editing. C.A. Warby: Data curation, investigation, methodology, writing–review and editing. J. Boehm: Conceptualization, supervision, investigation, methodology, writing–review and editing. S. Hardikar: Conceptualization, data curation, supervision, investigation, writing–review and editing. A. Ashworth: Resources, data curation, supervision, writing–review and editing. M. Schneider: Resources, data curation, supervision, writing–review and editing. A. Ulrich: Resources, supervision, writing–review and editing. P. Schrotz-King: Resources, supervision, writing–review and editing. J.C. Figueiredo: Resources, writing–review and editing. C.I. Li: Conceptualization, writing–review and editing. D. Shibata: Conceptualization, writing–review and editing. E.M. Siegel: Conceptualization, writing–review and editing. A.T. Toriola: Conceptualization, writing–review and editing. C.M. Ulrich: Conceptualization, supervision, investigation, writing–review and editing. K.L. Syrjala: Conceptualization, supervision, investigation, methodology, writing–review and editing. J. Ose: Supervision, investigation, methodology, writing–review and editing.

We thank the patients whose invaluable long-term contributions to the project make discoveries possible that improve care for future patients. We thank our collaborators on the ColoCare recruitment, particularly Hermann Brenner, Jenny Chang-Claude, and Michael Hoffmeister. We are grateful to all the study staff who have made this study possible: (i) from Heidelberg: Torsten K€olsch, Susanne Jakob, Clare Abbenhardt, Werner Diehl, Rifraz Farook, Lin Zielske, Anett Brendel, Marita Wenzel, Renate Skatula, and (ii) from Salt Lake City: Karen Salas, Sarah Fischbuch, Tyler Farr, and Anjelica Ashworth. J. Ose, B. Gigic, S. Hardikar, T. Lin, C. Himbert, C.A. Warby, J. Boehm, J.C. Figueiredo, A.T. Toriola, E.M. Siegel, C.I. Li, A. Ulrich, M. Schneider, D. Shibata, and C.M. Ulrich received funding by the NIH U01CA206110. J. Ose received funding by the NIH R03CA270473. C.M. Ulrich received funding from P30CA042014. C.M. Ulrich, J. Ose, and C.I. Li received funding from NIH R01CA207371 and R01CA189184. C.M. Ulrich and J. Ose additionally received funding from NIH R01CA254108. K. Syrjala received funding from NIH/NCI R01CA160684, R01CA215134, and U01CA246659. C.L. Lindley received funding from the University of Utah, Office of Undergraduate Research.

The publication costs of this article were defrayed in part by the payment of publication fees. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734.