Novel Methylation-Based Platform Shows Promise for Multicancer Early Detection

A genome-wide methylome enrichment platform demonstrated efficacy in detecting early-stage, low-shedding cancers with limited circulating tumor DNA (ctDNA), according to a poster presentation at the 2023 AACR Annual Meeting. Investigators believe that genome-wide methylome enrichment platforms may have clinical utility in the early detection of multiple cancers, as well as in identifying cancers that may have a better response to treatment.

For all-stage cancers, across 12 different cancer subtypes (n = 1232), the plasma cell-free DNA (cfDNA) test demonstrated an area under the curve (AUC) value of 0.94 (95% CI, 0.93-0.95). The AUC values for stages I (n = 284), II (n = 280), III (n = 327), and IV (n = 221), respectively, were 0.92 (95% CI, 0.90-0.94), 0.95 (95% CI, 0.93-0.96), 0.95 (95% CI, 0.94-0.96), and 0.97 (95% CI, 0.96-0.98). When each cancer type of was evaluated separately, the range of AUCs extended from 0.89 to 0.99.

“Initial training in this retrospective case-control study demonstrate feasibility of a genome-wide methylome enrichment platform for multicancer early detection, with promising differentiation of cancer cases and controls,” Ben H. Park, MD, PhD, of the Vanderbilt-Ingram Cancer Center, and co-investigators, wrote in the poster.

According to study authors, cfDNA tests are a promising approach in detecting early-stage cancers. Methylation-based tests are well-designed for this purpose, but different methodologies vary in performance. The purpose of this study was to develop a novel, genome-wide methylome enrichment platform for detecting multiple cancer types.

The design was retrospective; samples were used from commercial, Ontario tumor banks, and University Health Network biobanks. Samples from individuals who had been diagnosed with cancer who had not yet received treatment were used. Disease subtypes included bladder, breast, colorectal, endometrial, esophageal, head and neck, hepatobiliary, lung, ovarian, pancreatic, prostate, and renal cancers. These samples were paired with control samples for age- and sex-matched individuals who had no known cancer diagnosis. For most controls, there was at least 12 months of confirmed cancer-free follow-up post sample collection—control samples were excluded for analysis if the individual was over age 75 years, or if they were known to have multiple comorbidities.

A bisulfite-free, nondegradative genome-wide DNA methylation enrichment platform was used to assess all samples. This platform used 5 to 10 ng of cfDNA extracted from the plasma. Cell-free methylated DNA immunoprecipitation and high-throughput sequencing (cfmeDIPseq) technique were used. To distinguish cancer cases from noncancer controls, investigators developed a machine learning classifier.

Of note, the AUCs increased with each cancer stage, yet were determined to be high for early-stage cancers. Across 12 different cancer types, this approach yielded an AUC of 0.90 or greater for all stage disease except prostate cancer; the AUC was 0.89 (95% CI, 0.88-0.91).

Low-Shedding Cancers

In a subset of cancers that are considered low shedding (n = 595; bladder, breast, endometrial, prostate, and renal), the AUC was 0.91 (95% CI, 0.89-0.93). Only 8% of patients with these low shedding cancers had stage IV disease.

In bladder cancer, the AUC was 0.95 for all-stage disease (n = 94; 95% CI, 0.93-0.97). For stage I disease (n = 42), the AUC was 0.94 (95% CI, 0.92-0.97). For stage II disease (n = 19), the AUC was 0.96 (95% CI, 0.93-0.98), and for stage III disease (n = 16), the AUC was 0.98 (95% CI, 0.97-1.00). Stage IV disease (n = 4) AUC was not available.

For breast cancer, the AUC was 0.92 for all-stage disease (n = 98; 95% CI, 0.89-0.96). For stage I disease (n = 32), the AUC was 0.86 (95% CI, 0.80-0.92). For stage II disease (n = 3) and stage IV disease (n = 4), the AUCs were not available. The AUC for stage III breast cancer (n = 42) was 0.95 (95% CI, 0.93-0.97).

For any-stage endometrial cancer (n = 62), the AUC was 0.90 (95% CI, 0.87-0.93). For stage I disease (n = 38) the AUC was 0.90 (95% CI, 0.86-0.93), for stage II disease (n = 8), the AUC was 0.91 (95% CI, 0.85-0.98). Stage III (n = 4) and stage IV (n = 0) AUCs were not available.

Finally, in the renal cancer population (n = 158), the AUC was 0.91 (95% CI, 0.88-0.94). For stage I disease (n = 64), the AUC was 0.89 (95% CI, 0.85-0.94). The stage II (n = 2) AUC was not available. The stage III (n = 24) and IV (n = 15) AUCs were 0.89 (95% CI, 0.81-0.97) and 0.98 (95% CI, 0.96-1.00), respectively.

“The high detection of cancers with limited ctDNA that are typically difficult to detect (early-stage, low-shedding) is well suited for multicancer early detection applications and supports the added utility of identifying cancers that may have a better response to treatment,” study authors wrote.

Remaining Cancer Subsets

For patients with colorectal cancer, methylation-based tests demonstrated a 0.98 AUC for all-stage disease (n = 169; 95% CI, 0.95-0.99). For stage I disease (n = 18), the AUC was 0.97 (95% CI, 0.96-0.98). For stage II disease (n = 75), the AUC was 0.98 (95% CI, 0.97-0.99). For stage III disease (n = 44), the AUC was 0.97 (95% CI, 0.96-0.99), and for stage IV disease (n = 32), the AUC was 0.99 (95% CI, 0.97-1.00).

For patients with esophageal cancer (n = 77), the any-grade AUC was 0.99 (95% CI, 0.98-1.00). For stages I (n = 9), II (n = 16), III (n = 34), and IV (n = 18), respectively, the AUCs were 0.97 (95% CI, 0.95-0.99), 0.97 (95% CI, 0.93-1.00), 1.0 (95% CI, 0.99-1.00), and 1.0 (95% CI, 1.00-1.00).

In all-stage population of patients with head and neck cancer (n = 92), the AUC was 0.96 (95% CI, 0.94-0.98). In the stage I population (n = 7), the AUC was 0.93 (95% CI, 0.86-1.00), for stage II (n = 17), the AUC was 0.93 (95% CI, 0.83-1.00), and for stage III (n = 23), the AUC was 0.96 (95% CI, 0.94-0.99). The stage IV (n = 45) AUC was 0.97 (95% CI, 0.96-0.99).

In hepatobiliary cancer, all-stage AUC (n = 88) was 0.99 (95% CI, 0.98-0.99). For stage I disease (n = 24), the AUC was 0.98 (95% CI, 0.97-0.99). For stage II disease (n = 18), the AUC was 0.98 (95% CI, 0.97-1.00), for stage III disease (n = 23), the AUC was 0.99 (95% CI, 0.99-1.00), and for stage IV (n = 13), the AUC was 1.0 (95% CI, 0.99-1.00).

In lung cancer, any stage AUC was 0.96 (n = 125; 95% CI, 0.94-0.97). For stage I (n = 43), the AUC was 0.94 (95% CI, 0.92-0.97). For stages II (n = 21), III (n = 25), and IV (n = 36) the AUCs were 0.96 (95% CI, 0.93-0.99), 0.96 (95% CI, 0.93-1.00), and 0.98 (95% CI, 0.96-0.99), respectively.

In ovarian cancer, the AUC for all stage disease (n = 54) was 0.97 (95% CI, 0.95-0.99). For stage I disease (n = 6), the AUC was 0.98 (95% CI, 0.97-1.00), for stage II (n = 7) the AUC was 1.0 (95% CI, 0.99-1.00), stage 3 (n = 32) was 0.97 (95% CI, 0.95-0.99), and stage IV (n = 6) was 0.99 (95% CI, 0.97-1.00).

For patients with pancreatic cancer, the any-stage AUC was 0.99 (n = 32; 95% CI, 0.98-1.00). The AUCs for stage I (n = 0) and stage II (n = 4) were not available. The stage III (n = 5) AUC was 0.96 (95% CI, 0.91-1.00) and the stage IV (n = 23) AUC was 1.0 (95% CI, 99-1.00).

Study Implications

According to the study authors, these findings may be relevant in other areas of cancer that require detection of very trace amounts of ctDNA. They cited minimal residual disease as another possible application of this technology.

Study authors acknowledged that the retrospective design served as a limitation of this research. However, the ongoing CAMPERR study (NCT05366881) will seek to confirm the findings of this approach in prospectively collected samples.

Editor’s Note: This study was supported by adela.

Reference

Park BH, Shen SY, Min J, et al. Development of a genome-wide methylome enrichment platform for multi-cancer early detection. Presented at: 2023 AACR Annual Meeting; April 14-19, 2023; Orlando, FL.