Cancer Treatment and Genetics: New Discoveries from the 100 Thousand Genomes Project

UK researchers integrated genomic and clinical data for 13,880 solid tumors (arising in organs or tissues) across 33 cancer types and identified genes that influence prognosis (the course of disease).

The researchers linked genomic and clinical data to identify genes affecting prognosis in patients with cancer. Their findings were reported on January 11, 2024, in Nature Medicine , where the study was published.

Linking Genomic and Clinical Data Helps Explain Cancer Outcomes

Demand is growing for molecular cancer tests that identify cancer susceptibility genes and pharmacogenomic markers to improve treatment effectiveness. These tests can determine an individual's predisposition to cancer and support personalized treatment.

Conducted within the UK National Health Service (NHS), the 100,000 Genomes Project was a major oncology initiative aiming to provide standardized, high-throughput whole-genome sequencing (WGS) for patients with cancer and rare diseases.

Design of the 100,000 Genomes Project

• The project uses a bioinformatics system (software pipeline) approved by the International Organization for Standardization to detect and characterize genetic variants (alterations).
• This system forms part of large-scale whole-genome sequencing (WGS) work used in cancer care within the National Health Service (NHS). Patients consented to their genomic data being combined with anonymized health records and shared in a secure research environment to advance knowledge of cancer. These data were securely linked to long-term real-world data to establish a national molecular data platform called the National Genomic Research Library.
• The library includes national clinical datasets such as cancer registries, systemic anticancer treatments, hospital records and mortality statistics. This approach allows findings from genomic research to feed back into genomic healthcare services, enabling research to contribute directly to clinical practice and treatment.

13,880 Tumor Samples Across 33 Cancer Types

For this study, researchers performed whole-genome sequencing (WGS) on 13,880 tumor samples across 33 cancer types and also examined 13,616 matched normal samples.

The most common tumor types were breast cancer (2925 patients), colon adenocarcinoma (1948), sarcoma (1617) and clear-cell kidney cancer (1163).

Clinically significant genetic variants were found in more than 50% of tumor samples from glioblastoma, low-grade glioma, melanoma, head and neck squamous cell carcinoma, colon and rectal adenocarcinoma, and lung adenocarcinoma.

Clinically significant variants were identified in 20% to 49% of some cancer types, including breast, ovarian, uterine and urinary tract cancers, lung squamous cell carcinoma, mesothelioma and sarcoma. In pancreatic, prostate, esophageal and stomach cancers, mutations in cancer genes important for treatment and diagnosis were found in fewer than 20%.

Somatic and Germline (Inherited) Variants

TP53 mutations were detected in 39% of the 5411 cancer samples examined, making this the most frequently mutated gene among the cancers studied. TP53 mutations were observed in more than 70% of cancers such as uterine serous endometrial carcinoma, ovarian high-grade serous carcinoma, lung squamous cell carcinoma, rectal adenocarcinoma, esophageal adenocarcinoma and esophageal squamous cell carcinoma.

PIK3CA was mutated in 19.8% of the 2750 patients examined, making it the second most frequently mutated gene. Mutations were particularly frequent in uterine endometrial carcinomas (53.5%), ovarian endometrioid adenocarcinomas (49%), breast carcinomas (42.2%), uterine serous endometrial carcinomas (38.1%) and colon adenocarcinomas (26.5%). These rates indicate a relatively high likelihood of PIK3CA mutations in these cancer types.

About 13% of patients diagnosed with high-grade serous ovarian carcinoma had inherited, potentially treatment-guiding germline BRCA1 and BRCA2 gene alterations. Patients carrying these changes were generally found to have developed cancer at a younger age.

Similarly, patients with germline changes in genes that repair DNA errors (mismatch repair genes) developed colon cancer at a younger age. These inherited changes can pass through families and influence cancer risk. In particular, patients carrying changes in genes involved in DNA repair (homologous recombination repair genes) who had high-grade serous ovarian carcinoma or breast cancer, and patients with clear-cell kidney cancer carrying VHL alterations, were also observed to develop cancer earlier.

Biomarkers and Mutational Signatures

The researchers found that tumor mutational burden (TMB) varied considerably across cancer types. Melanoma and lung adenocarcinoma, for example, had a higher average TMB than other cancers.

Specific genetic changes (mutational signatures) were also associated with particular cancers. APOBEC signatures, for example, were associated with breast cancer, head and neck squamous cell carcinoma, urinary tract cancers and lung adenocarcinoma. Smoking-associated genetic changes were generally linked to lung cancers, while those associated with ultraviolet exposure were linked particularly to melanoma.

Homologous recombination deficiency (HRD), a genetic condition, was most prevalent in high-grade serous ovarian carcinoma, occurring in 40% of cases. However, HRD was also observed in other tumor types. It may be an important indicator for treatment and risk assessment.

What Do These Genetic Data Mean for Clinical Practice?

By linking whole-genome sequencing (WGS) data with patients' long-term health records, researchers examined treatment outcomes based on markers across the genome.

For example, among 189 patients receiving platinum-based chemotherapy, deficiencies in particular DNA repair genes (HRD) significantly reduced the risk of death. The hazard ratio (HR) was 0.37, with a 95% confidence interval of 0.23 to 0.61, a statistically significant result (P <0.001). This was particularly evident in patients with breast cancer or high-grade serous ovarian cancer.

Among patients with melanoma, those with low TMB had shorter overall survival than those with high TMB. The hazard ratio was 2.34, with a 95% confidence interval of 1.14 to 4.80, a statistically significant difference (P =0.015). No similar effect was observed in lung cancer, where the result was not statistically significant (P =0.72). These findings show how treatment response can differ by cancer type and genetic profile, underscoring the importance of personalized treatment.

This suggests that TMB plays an important role in cancer prognosis, as earlier research has also highlighted. More detailed investigation of genetic markers (pangenomic biomarkers) is needed both to predict overall survival and to anticipate response to immunotherapy.

15 Different Genes Affecting Overall Survival Identified

Among these genes, CDKN2A had the greatest effect on overall survival according to the findings (hazard ratio [HR] 2.3; 95% confidence interval [CI] 2.0-2.6; P <1×10−10). The researchers note that CDKN2A mutation is associated with high-grade disease and poor prognosis in certain cancer subtypes, particularly glioma and soft tissue sarcoma. This indicates that CDKN2A is an important prognostic marker in these cancers.

As expected, the research found that mutations in KRAS were associated with poor prognosis in colorectal cancer, while mutations in both KRAS and TP53 were associated with poor prognosis in non-small cell lung cancer. PIK3CA mutations, by contrast, were associated with favorable outcomes, meaning that cancers carrying these mutations had a better prognosis.

The researchers expect WGS data to provide further insights into cancer prognosis as the dataset grows, particularly through integration with longitudinal patient data and other sequencing approaches, such as cell-free DNA (cfDNA) and single-cell sequencing.

In conclusion, the researchers emphasize that integrating genomic data with real-world clinical data made it possible to analyze cancer survival and better understand how cancer genetics affects patient outcomes. This approach is an important step in cancer treatment and research.