Statistics from Altmetric.com
We read with interest the work by Gündert and colleagues, which describes a new prognostic classifier for non-metastatic colorectal cancer (CRC), ProMCol, derived from DNA methylation levels at 20 CpG sites in colorectal tumour tissue.1 Here, we tested the ProMCol classifier on an independent cohort of 526 non-metastatic CRC tumours from the Melbourne Collaborative Cohort Study (MCCS). The MCCS is a prospective cohort study of 41 513 healthy adult volunteers recruited between 1990 and 1994.2 By 31 December 2009, 1046 participants had a first histopathological diagnosis of invasive adenocarcinoma of the colon or rectum identified through the Victorian Cancer Registry (VCR) following the baseline study visit (a total of 1101 CRCs). Characterisation of this CRC-affected cohort is described in Buchanan et al.3 Vital status was ascertained through the VCR and the National Death Index.
From all 1046 participants with a CRC diagnosis, we excluded 26 tumours with the tumour site listed as overlapping lesions of rectum, anus and anal canal (International Statistical Classification of Diseases and Related Health Problems 10th Revision code C21.8) or had unknown site, we removed 228 tumours without available formalin-fixed paraffin embedded (FFPE) tissue specimens, a further 142 tumours were unavailable for DNA methylation testing, and finally we removed 124 metastatic cases (stage 4 American Joint Committee on Cancer (AJCC)) and tumours with unknown stages, leaving a total of 526 CRC tumours for DNA methylation analysis. Genome-wide DNA methylation profiles were assessed using the Inifnium HM450K on macrodissected FFPE tumour samples as previously described.4 Raw HM450K data were imported into R programming software V.3.3.2 and processed using the minfi bioconductor package.5 The data underwent the standard Illumina and SWAN normalisation.6 The ProMCol classifier for each tumour sample was calculated as indicated in Gündert et al,1 where β-values from each of 20 ProMCol-related CpG sites were first obtained then multiplied by the Principal Component Analysis (PCA) weights for individual CpG sites.
In the 526 MCCS CRC tumours, the ProMCol values ranged from 1.67 to 3.59 with a median of 2.86 compared with a median of 2.86 and a range of 1.20 to 3.51 described in Gündert et al. As per the original study,1 we divided our tumour samples into high (>2.86) and low (≤2.86) ProMCol groups using the median as cut-off. Overall survival was assessed using Kaplan-Meier curves. We used Cox regression,7 with time since CRC diagnosis as the time axis, to estimate HRs and 95% CIs for overall survival. All multivariate models were adjusted for age, sex, country of birth, smoking status, tumour mismatch repair status determined by immunohistochemistry, anatomical location of tumour and AJCC stages 1–3. We examined each model for outliers and influential points and used Schoenfeld residuals to assess the proportional hazard assumption; there was no evidence that it was violated. All statistical analyses were performed using Stata V.14.1 (StataCorp).
The mean age at CRC diagnosis for the 526 MCCS participants studied was 68.6 years (SD=8.20, range: 44.6–85.2) compared with 69.3 (range: 33–94) in the study by Gündert et al. During an average follow-up of 11.0 years, a total of 275 deaths occurred. The overall survival was higher for the upper ProMCol group compared with the lower ProMCol group (HR=0.79, 95% CI: 0.70 to 0.88, p<0.01) (table 1). The 5-year overall survival was 84% (95% CI: 63% to 74%) for the high ProMCol group and 69% (95% CI: 79% to 88%) for the low ProMCol group (figure 1). By fitting a model with ProMCol as a quartile increment, we also found significant associations with age (p<0.01), smoking status (p=0.03) and AJCC stage (3 vs 1, p<0.01). Overall, our results were consistent with those reported by Gündert et al, where the higher ProMCol value was associated with a better prognosis in individuals with non-metastatic CRC.
We have confirmed in an independent cohort the use of the ProMCol classifier in predicting overall survival for patients with non-metastatic CRC. The ProMCol classifier used in conjunction with other commonly used clinicopathological characteristics might provide a more accurate prognosis for CRC patients with non-metastatic disease, enabling more predictable treatment options, preventing over and undertreatment. Although the ProMCol classifier was defined using the Infinium Human Methylation 450K, developing an accurate and cost-effective diagnostic technique targeting the set of relevant CpG sites should be prioritised.
We would like to thank all participants of the Melbourne Collaborative Cohort Study, the original investigators and all staff who have been involved recruiting participants and working on follow-up.
MCS and DDB contributed equally.
Contributors This study was first conceived by JEJ, HJ and DDB. EMW and JEJ performed laboratory experiments. JEJ and HJ performed bioinformatics and statistical analysis. HJ, DRE, RLM, GGG and DDB faciliated the inclusion of the data from the MCCS. MC and CR provided molecular characterisation and paghological review of FFPE slides, respectively. The manuscript was first constructed by JEJ, HJ and DDB. MC, CR, IMW, MAJ, JLH, DRE, RLM, GGG and MCS provided critical reviews of the analysis and the manuscript. All authors reviewed and approved the final manuscript.
Funding This project was supported by grant from the National Health and Medical Research Council (NHMRC) (grant 1027505). MCCS recruitment was funded by VicHealth and Cancer Council Victoria. The MCCS was further supported by Australian NHMRC grants 209057, 251553 and 504711 and by infrastructure provided by Cancer Council Victoria. Cases were ascertained through the Victorian Cancer Registry (VCR) and the Australian Cancer Database (Australian Institute of Health and Welfare). HJ is supported by a Victorian Cancer Agency Early Career Seed Grant Fellowship. MCS is a Senior Research Fellow and JLH is a Senior Principle Research Fellow of the NHMRC. DDB is a NHMRC RD Wright Biomedical Fellow.
Disclaimer Authors had full responsibility for the design of the study, the collection of the data, the analysis and interpretaion of the data, the decision to submit the manuscript for publication and the writing of the manuscript.
Competing interests None declared.
Patient consent Detail has been removed from this case description/these case descriptions to ensure anonymity. The editors and reviewers have seen the detailed information available and are satisfied that the information backs up the case the authors are making.
Ethics approval This study was approved by the Cancer Council Victoria Human Research Ethics Committee (IEC No. 9001).
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.