Decoding Cancer Without a Biopsy

Molecular profiling of tumors has brought a new dawn in cancer diagnostics and treatment, however, the realization of this promise often meet a significant roadblock which is obtaining an adequate amount of tumor tissue. For cancers like ovarian cancer, where obtaining biopsy is invasive and risky, or biliary tract cancer, where samples are often inadequate, biopsy-based diagnosis fall short. This limitation could be curbed using a non-invasive alternative with remarkable potential in the form of circulating tumor DNA (ctDNA).

ctDNA are short, double-stranded DNA fragments, typically 40 to 200 base pairs long, originate from either nuclear or mitochondrial DNA and carry a mix of coding and non-coding sequences. ctDNA is not confined to blood; it can be secreted in nearly all types of human body fluids — urine, saliva, cerebrospinal fluid (CSF), ascites, pleural effusion, and even pericardial effusion.

The field of ctDNA is still in infancy where different terms often prove to be a matter of confusion. For example, terms like ctDNA (circulating tumor DNA) and cfDNA (cell free DNA). cfDNA is DNA released by differnt cell types (WBCs, endothelial cells), but ctDNA is DNA released by tumor cells. In asymptomatic individual, ctDNA constitutes only 0.1% of the total cfDNA and screening them would require ultra-deep sequencing (>10,000x coverage) with the incorporation of unique molecular identifiers (UMIs) to reduce sequencing errors, which is currently prohibitively expensive.

ctDNA as the New Frontier

In a study analyzing ctDNA alongside biopsy, a concordance rate of approximately 84.8% (95% CI: 78.5–89.5%) was observed, highlighting its reliability as a surrogate for tissue biopsies. Furthermore, a strong correlation was noted between ctDNA and tissue variant allele frequencies (VAFs), with a coefficient of 0.62 (p = 1.0 × 10⁻²⁰).

Three Key Approaches for Cancer Detection Using cfDNA

Somatic Mutation Analysis This method focuses on identifying tumor-specific genetic mutations, such as point mutations, insertions, deletions, and copy number variations. While highly informative, this technique is labor-intensive, requiring deep sequencing to pinpoint these alterations, making it costly and time-consuming.

Methylation Profiling Tumors exhibit unique DNA methylation patterns that distinguish them from healthy tissue. This approach examines these epigenetic modifications in cfDNA, enabling cancer detection and tissue-of-origin determination. Though powerful, it involves complex workflows and high costs, limiting its scalability for routine use.

Whole-Genome Fragmentation Profiles A more recent and cost-effective method, this approach analyzes the fragmentation patterns of cfDNA across the entire genome. Tumor-derived cfDNA exhibits distinct fragmentation profiles due to its release mechanisms and cellular context. This technique requires less intensive sequencing and offers a faster, scalable alternative to somatic mutation and methylation-based analyses.

Q: Can the source of mutations (cfDNA/ctDNA) be distinguished?

Differentiating between cfDNA derived from white blood cells (WBCs) and ctDNA is a complex challenge. One proposed approach involves studying nucleosome positioning, which correlates with the epigenetic characteristics of each cell type. This method can provide clues about the tissue of origin for cfDNA.

ctDNA: A Prognostic Marker

ctDNA levels directly correspond to the disease burden. Patients with higher ctDNA levels often face poorer survival outcomes compared to those with lower levels. For clinicians, this connection is invaluable. But ctDNA’s utility doesn’t stop there. It also acts as a real-time monitor of treatment response. When therapies work effectively, ctDNA levels drop, signaling tumor shrinkage and a positive response. Conversely, stable or rising ctDNA levels can warn of treatment resistance or disease progression long before imaging tests reveal the full picture.

In recent years, ctDNA has gained recognition as a marker of Minimal Residual Disease (MRD). Even after successful treatment, small numbers of cancer cells can linger undetected, poised to cause a recurrence. Here, ctDNA shines by identifying these remnants at their earliest, subclinical stage, allowing clinicians to anticipate relapses and intervene preemptively.

This predictive power has been vividly demonstrated in studies on pancreatic adenocarcinoma. Researchers found that detectable levels of ctDNA after surgery reliably forecast clinical recurrence, often months before traditional methods could confirm it. For patients and doctors, this is a game-changer — it means early warnings, better preparedness, and a chance to act decisively.

Correlations Between cfDNA Levels, Tumor Burden, and Staging

In cancer patients, cfDNA levels are approximately ten times higher than in healthy individuals. Research has consistently shown that cfDNA levels correlate with tumor size and disease staging. A striking example comes from a study involving 49 patients with large B-cell lymphoma. Pre- and post-therapy analyses of cfDNA levels revealed a clear pattern:

• Responders: Patients who achieved complete responses after therapy showed a significant decrease in cfDNA concentrations.
• Non-responders: Those with persistent or progressive disease exhibited increased cfDNA levels.

Moreover, the variant allele frequency (VAF) of cfDNA, representing the proportion of tumor-derived mutations, was higher in advanced stages of the disease. This underscores the potential of cfDNA not only to reflect disease burden but also to provide insights into its aggressiveness.

Q: Can the mutations in cfDNA be missed?

Interestingly, in cases where specific mutations were missed in cfDNA, the median VAF in tissue was significantly lower than in cases where these mutations were detected in cfDNA.

• Median tissue VAF for undetected mutations: 20.94% (IQR: 12.15–27.05%)
• Median tissue VAF for detected mutations: 32.20% (IQR: 24.66–48.10%)

Challenges into ctDNA Detection

Tumors located in the central nervous system (CNS), such as gliomas, are particularly challenging. These tumors release minimal amounts of ctDNA into the bloodstream, making it nearly undetectable in glioma patients. This limitation extends to other cancers like medulloblastomas, kidney, prostate, and thyroid cancers, where ctDNA is harder to detect due to its low presence or specific release patterns.

However, ctDNA detection is much more successful in certain cancers, especially those where the tumors shed a higher concentration of ctDNA into surrounding body fluids. For example, in ovarian cancer, the use of peritoneal washings during surgery has proven to be an effective way to collect ctDNA, providing valuable molecular insights. Similarly, ascites (fluid buildup in the abdomen) from ovarian cancer patients has been shown to harbor cfDNA, offering clinically actionable information that matches tissue biopsy results. This allows for opportunistic molecular profiling, enabling better-informed treatment decisions even in the absence of a traditional biopsy.

For advanced-stage cancers like ovarian, liver, pancreas, bladder, colon, lung, stomach, breast, esophagus, and head and neck cancers, ctDNA is more easily detectable. Similarly, cancers like neuroblastoma and melanoma also present higher levels of ctDNA in the bloodstream, facilitating detection and molecular characterization.

Interestingly, the histology of a tumor plays a significant role in ctDNA levels. Squamous cell carcinomas and triple-negative breast cancer (TNBC), for instance, tend to shed higher percentages of ctDNA into circulation.

These findings underscore the importance of understanding the nuances of ctDNA dynamics across different cancers. Circulating free DNA (cfDNA) might be small in size, but its role in understanding the human body is monumental.

References:

1. https://doi.org/10.1016/j.jhep.2024.10.020
2. https://pmc.ncbi.nlm.nih.gov/articles/PMC10496721/
3. https://www.frontiersin.org/journals/oncology/articles/10.3389/fonc.2022.943253/full
4. https://ascopubs.org/doi/10.1200/JCO.2024.42.16_suppl.107