Lung cancer leads in mortality ahead of all other cancer deaths in both developed and undeveloped countries and is the second most commonly diagnosed cancer after breast cancer. While the disease is showing nominal decline due to successful public health measures (like cessation programs) and people abstaining from smoke and tobacco products, lung and bronchial cancer is relentless with 5-year survival rates at only 20% taking an estimated 135,000 lives in the US this year.
The challenge with lung cancer lies within diagnosis, response to treatment and detection of disease progression. While most instances of lung cancer are found in smokers, a study published in April 2019 found that lung cancer in people who have never smoked is showing in 20% of cases due to second-hand smoke, occupational carcinogen exposure, and outdoor pollution.
With respect to all cases, however, most symptomatic and diagnosed patients are late stage where the cancer has already metastasized.
Methods detecting early stage cancer are aimed at screening smokers and nonsmokers over the age of 55 where medical imaging allows clinicians to non-invasively diagnosis and monitor a patient’s tumor. For these at risk candidates, chest CT scans give the greatest decrease in mortality compared to chest x-ray or sputum cytology.
- False-negative test results can occur.
- False-positive test results can occur.
- Chest x-rays and low-dose spiral CT scans expose the chest to radiation.
Occurrences like these are going to lead to either delayed diagnosis or costly, unnecessary and invasive biopsy procedures that can lead to further complications. A lung cancer biopsy can cause part of the lung to collapse requiring potential need for surgery while low-dose spiral CT may increase the risk of cancer.
Once that cancer diagnosis is made, identifying the severity of the disease and monitoring lesions during and after treatment pose additional challenges. Throughout treatment, pathologists perform manual, tedious microscopic examination of subtle histopathological patterns in highly complex tissue images. Quantitative assessments like these are subjective and reliability is unfortunately dependent on inter- and intra-observer variation.
In efforts to change the landscape of lung cancer detection and prognosis, researches are looking at new methods to improve sample collection, early diagnosis and efficient monitoring of tumors.
The Lung Cancer Institute at Johnson & Johnson is working on developing a number of biomarkers with application in the diagnosis of pulmonary nodules. Monarch, a robotic bronchoscopy platform carried by Ethicon and Auris Health, part of the Johnson & Johnson Medical Devices Companies, allows clinicians to sample lesions through airways and potentially treat, locally, on the same path as well.
Because one of the major hallmarks of cancer, splicing deregulation, affects progression, metastasis and therapy resistance, Janssen Pharmaceutical Companies of Johnson & Johnson and Artificial Intelligence (AI) will also play a critical role in precision oncology.
Artificial intelligence, especially machine and deep learning, has been validated in biomedical research, and can be applied in all stages of drug discovery. There is great potential in pathology image analysis such as target validation, identification of prognostic biomarkers, metastasis detection and central neural networks (CNN) and is quite attractive to an institute like LCI.
In April of this year, Envisagenics entered into a research program agreement with the Lung Cancer Institute at Johnson & Johnson. Envisagenics has long believed that cancer-associated splicing deregulation may be a novel source of clinically applicable biomarkers. I met with Maria Luisa Pineda, Ph.D., co-founder and Chief Executive Officer of New York based Envisagenics, in 2019 at a WIB panel event (From the Bench to the Boardroom) where she talked about the importance of partnerships with champions in pharma that are interested in novel technologies. With the right collaboration partners, the company is able to accelerate R&D to develop RNA therapeutics against genetic and complex immune-oncology (IO) diseases.
Their AI cloud-based software platform, SpliceCore®, integrates machine learning (ML) algorithms with high performance computing algorithms to analyze their proprietary database of more than 5 million potential RNA splicing errors. Detection of these RNA splicing events can “identify early determinants of lung cancer risk” and validate potential drug targets faster than traditional methods. According to Martin Akerman, Ph.D., co-founder and CTO, “SpliceCore can extract these biological insights because it can scan through millions of potential RNA splicing events from 1,000 patient samples in two hours.”
Envisagencis spun out of Cold Spring Harbors Lab with validated ML algorithms that successfully identified RNA splicing errors in children diagnosed with spinal muscular atrophy. Following these discoveries, members of the same team developed the 2016 FDA-approved therapeutic, Spinraza- the only drug cleared at that time to treat the rare neuromuscular disorder.
As a resident of JLABS NYC, the company is also focused on potential targets for in-house therapeutic development in genetic diseases like breast cancer, leukemia and ALS. Their biotechnology differentiates itself from “mainstream transcriptomics” by examining exons rather than genes, which is a more informative unit in both pathology and immunity. In 2019, the company received their third SBIR grant from the NCI to enhance their SpliceCore platform to conduct pre-clinical studies developing ENV-0205 for the treatment of triple negative breast cancer (TNBC). In their therapeutic asset pipeline, the lead compound targets a novel isoform that is present in 65% of patients suffering from this aggressive type of breast cancer.
With this grant, the company will also collaborate with expert researchers at Memorial Sloan Kettering Cancer Center to study myeloid leukemia. Researchers at Envisagenics predecessor labs, Cold Springs Harbor Labs (CSHL), in collaboration with Memorial Sloan Kettering Cancer Center (MSKCC), revealed earlier this year their discovery of unique splicing errors in blood cancer. Transcription error was identified in a process called Nonsense-mediated mRNA decay (NMD). They saw that when gene SRSF2 mutated, NMD destroys not only the messages that contained mistakes, but also the healthy proteins. A proof of concept study at the German Center for Neurodegenerative Diseases (DZNE) and the University of Bonn applied ML to detect the blood disease, acute myeloid leukemia (AML) and showed the hit rate to predict the onset of AML at above 99% accuracy. These results were based on the analysis of the gene activity of cells found in the blood. Envisagenic’s exon-centric approach to RNA-seq and SpliceCore’s software has the potential to uncover transcriptomic dynamics, invisible to bulk- and gene-level analysis, to go beyond current proof of concept models. With the work Dr. Pineda’s team is performing in lung, breast and blood cancer, the company can expect, in the next 18 months to advance closer to development of disease-specific splicing variants that will function as therapeutic targets.