Genomic alterations may be used to predict whether carcinoma in situ will develop into an invasive tumor, according to a recent study in Nature Medicine examining human lung tissue samples.

The new research suggests that “many changes observed in cancer are present much earlier in the disease process than previously thought,” said Sam Janes, PhD, a professor of respiratory medicine at the University College London in the United Kingdom, who coauthored the study.

In the study, researchers collected 129 carcinoma in situ (CIS) biopsy samples from the lung tissues of 85 people. The investigators then performed genomic analyses of the samples, including gene-expression profiling, methylation profiling, and whole-genome sequencing. The researchers subsequently followed the study participants for more than 5 years to see which people developed lung squamous cell carcinoma.

Continue Reading

The researchers took a close look at 29 samples collected from people whose CIS developed into lung squamous cell carcinoma and 10 samples harvested from people whose CIS regressed. The scientists found that the former samples had a much greater number of mutations, and more changes in the number of copies of certain genes, compared with the regressive samples.

Related Articles

Moreover, nearly all of the progressive samples had mutations in the tumor-suppressor gene TP53. The progressive samples also had certain chromosomal aberrations that are typically present in squamous-cell carcinomas. Somewhat unexpectedly, 3 out of 4 CIS growths with TP53 mutations and alterations that looked regressive per visual monitoring eventually progressed.

By means of statistical analysis, the researchers also concluded that the DNA-methylation patterns in the majority of the regressive CIS samples looked more like the patterns of healthy lung cells, compared with the patterns present in progressive tumors. The only exceptions were the samples that progressed after initially being categorized as regressive. The researchers were able to differentiate between regressive and progressive samples using either DNA methylation data or copy number data.

The researchers also used the data to predict whether CIS would develop into an invasive tumor. The investigators tested their approach on CIS biopsies that had previously been set aside for this purpose, and then tested it again on data from human lung samples that came from The Cancer Genome Atlas (TCGA) project. The researchers found that they were able to predict the probability of CIS progression with near-perfect accuracy.

Richard Lazzaro, MD, chief of thoracic surgery at Lenox Hill Hospital in New York City, who was not involved in the research, called the new results “exciting.” Even more exciting would be a potential application of the findings to developing treatments that could address the genetic abnormalities leading to invasive cancer, and that could stop the disease from progressing, Dr Lazzaro added.

Still, more work is needed before potential practical applications are developed. “While we anticipate that this work is certainly pushing forward our understanding of cancer precursor biology and might improve early detection, reduce overtreatment, for example, by fostering [preventive] therapies targeting early clonal events in lung cancer, one should be aware that most of the technical approaches used in the study are not part of the typical toolbox in routine clinical practice,” said Luca Quagliata, PhD, global head of medical affairs, clinical next-generation sequencing and oncology at Thermo Fisher Scientific in Zug, Switzerland, who was not involved in the study. “Thus, despite the invaluable importance of these data, before patients will be able to benefit from it, a more agile technical approach needs to be implemented to derive the same set of info.”


  1. Teixeira VH, Pipinikas CP, Pennycuick A, et al. Deciphering the genomic, epigenomic, and transcriptomic landscapes of pre-invasive lung cancer lesions. Nat Med. 2019;35(3):517-525.