Cancer is essentially a disease of the genome which evolves with the accumulation of somatic mutations.1 Explosive advances in next-generation sequencing (NGS) have greatly improved the ability to identify actionable cancer mutations, both for solid and hematological malignancies, and sparked a new era of oncology care.2 But accurate analysis and proper interpretation of the complex genomic data produced by NGS remain key hurdles.

Spearheading the revolution for genome-led cancer care, SOPHiA GENETICS has developed SOPHiA™ artificial intelligence, the advanced technology for Data-Driven Medicine. SOPHiA accurately analyzes large amounts of genomic data and helps experts quickly detect and characterize actionable mutations in solid and blood cancers.


Finding the signal in the noise

High-throughput genomic data analysis is rife with noises and biases that can be introduced in any step of the NGS workflow. SOPHiA has been exposed to many NGS-based genomic data, from different sample types, enrichment kits and sequencing platforms, produced by hundreds of healthcare institutions. This exposition allowed us to understand all the biases that experts may encounter while analyzing datasets.

Because the more input the AI receives, the smarter it becomes, we have fed SOPHiA with datasets from real clinical samples containing over half a million distinct confirmed variants. We have trained SOPHiA to improve its own ability to accurately detect and characterize genomic alterations and achieve excellent analytical performance, regardless of the type of sample or sequencer used, making SOPHiA a universal technology.


More than variant detection

SOPHiA has been implemented to address the challenges of interpreting the daunting number of genomic variants produced by NGS, ensuring complete decision support for solid tumors and hematological malignancies.

The detection of mutations in cancer is more challenging due to very low variant frequency and mixed clonal cell populations.3

The SOPHiA AI platform for oncology offers a comprehensive workflow, enabling experts to easily visualize, interpret and report mutations associated with cancer and available treatment. Its hotspot screening feature eases the visu­alization of the actionable mutations. With pre-classified variants and customized variant filtering options, experts can easily accelerate the data interpretation process.

The platform integrates the OncoPortal, a decision support functionality based on precision medicine intelligence. SOPHiA matches genomic alterations with curated databases of evi­dence-based clinical associations. Such associations encom­pass a combination of genomic alterations, cancer types and therapies. This information highlights the ac­tionability of the tumor profile within the same and in other tumor entities. It also uses inclusion and exclusion criteria to identify clinical trials, both locally and at the global level.


Case study: How to optimize management of lung cancer?

With nearly 1.6 million deaths reported each year, lung cancer is the leading cancer killer both in men and women globally. Despite a reduction in lung cancer mortality in the US in the past 20 years, lung cancer incidence is increasing in the developing world due to a rise in tobacco smoking in many countries.4

There are 2 main types of lung cancer:

  • Approximately 80% to 85% of lung cancers are non-small cell lung cancer (NSCLC)
  • Approximately 10% to 15% are small cell lung cancer (SCLC)5

According to the American Lung Association, the lung cancer five-year survival rate (18.6 percent) is lower than many other cancer sites, such as colorectal (64.5 percent), breast (89.6 percent) and prostate (98.2 percent).6

Until recently, there were no proven methods to detect early stage lung cancer. In the vast majority of the cases, the disease is discovered at an advanced stage, which is not amenable to curative therapies.

In 2003, Gefitinib, an oral epidermal growth factor receptor (EGFR) tyrosine kinase inhibitor (TKI), received FDA approval for the treatment of advanced-stage NSCLC with an objective tumor response rate of 10%–19%. Understanding the biology behind these differential responses led to the discovery of two sensitizing EGFR mutations (exon 19 deletion [del 19] and exon 21 L858R point mutation) that were present in subjects with a favorable response to TKI therapy.7

SOPHiA AI provides very high performance in detecting acquired resistance mechanisms, such as p.T790M and p.C797S mutations in EGFR gene using first to third generation inhibitors, and accurately identifies molecular targets for optimized management of NSCLC.


Radiomics: a promising approach towards precision oncology

Advances in high-throughput molecular technologies hold great promise for the development of precision oncology. Similarly, medical imaging technologies help experts measure the properties of the cancer tissues such as size, shape and texture. Identifying properties of tumors through medical images is part of the standard of care. As image interpretation can be subjective, the development of robust and standardized radiomics technology is required to analyze disease progression. Quantitative image analyses provide a comprehensive representation of the tumor.


SOPHiA Radiomics: the combined radiomic-genomic solution for precision oncology

Cancer is influenced by many parameters. Its behavior can’t be predicted based on histopathological analysis alone.

The complementary relationship between molecular profiling and medical imaging strongly suggests that combining the use of molecular and imaging biomarkers will significantly improve cancer care. It provides insight into how molecular mechanisms and imaging phenotypes can have a major impact on treatment plan.8

As an example, for NSCLC, which is the leading cause of cancer death9, radiomics would allow experts to investigate the relationships between image features, tumor molecular phenotype, and survival outcomes, and would help optimize the standard of care of this life-threatening illness. 

SOPHiA GENETICS has developed SOPHiA Radiomics, a decision support technology that analyzes medical images and combines them with biological data to deliver a more accurate disease prognosis. This powerful radiomics solution is designed to go beyond the usual RECIST and PERCIST criteria, to deliver finer biomarkers and enhance cancer care.



1: Macconaill LEGarraway LA. Clinical implications of the cancer genome. J Clin Oncol. 2010 Dec 10; 28(35):5219-28. doi: 10.1200/JCO.2009.27.4944. Epub 2010 Oct 25.

2: Nakagawa HFujita M. Whole genome sequencing analysis for cancer genomics and precision medicine. Cancer Sci. 2018 Mar;109(3):513-522. doi: 10.1111/cas.13505. Epub 2018 Feb 26.


3: Nadeu F. et al. Clinical impact of clonal and subclonal TP53, SF3B1, BIRC3, NOTCH1, and ATM mutations in chronic lymphocytic leukemia. Blood. 2016 Apr 28; 127(17): 2122–2130.


4: National Lung Screening Trial Research Team, et al. Reduced lung-cancer mortality with low-dose computed tomographic screening. N Engl J Med. 2011;365(5):395–409.


5: Small%20Cell%20Lung%20Cancer?


6: U.S. National Institute of Health, National Cancer Institute. SEER Cancer Statistics Review, 1975–2015


7: Paez JG, et al. EGFR mutations in lung cancer: correlation with clinical response to gefitinib therapy. Science. 2004;304(5676):1497–1500.

8 : Bakr S. A radiogenomic dataset of non-small cell lung cancer. Sci Data. 2018 Oct 16;5:180202. doi: 10.1038/sdata.2018.202.


9: Jemal A. Cancer statistics, 2010. CA Cancer J Clin. 2010 Sep-Oct;60(5):277-300. doi: 10.3322/caac.20073. Epub 2010 Jul 7.